{ "license": "CC BY 4.0 — https://creativecommons.org/licenses/by/4.0/", "attribution": "VOLO (flyvolo.ai) — please credit and link to https://flyvolo.ai/en/data", "notes": [ "The impact index is not a probability of losing a job; it never appears without its confidence and assessment date.", "Task direction, evidence stage and the impact index are separate readings and must not be summed or combined.", "Every task carries what its judgement does not establish; carry it with the claim." ], "occupations": [ { "slug": "accountant", "name": "Accountant / Bookkeeper", "sector": "business", "impact_index": 68, "impact_index_confidence": "medium", "assessed_on": "2026-09-09", "onet_soc": "13-2011.00 (broader); 43-3031.00 (exact)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/accountant" }, { "slug": "customer-service-representative", "name": "Customer service representative", "sector": "retail", "impact_index": 74, "impact_index_confidence": "medium", "assessed_on": "2026-09-09", "onet_soc": "43-4051.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/customer-service-representative" }, { "slug": "translator", "name": "Translator / Interpreter", "sector": "creative", "impact_index": 79, "impact_index_confidence": "medium", "assessed_on": "2026-09-09", "onet_soc": "27-3091.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/translator" }, { "slug": "truck-driver", "name": "Truck driver", "sector": "transport", "impact_index": 45, "impact_index_confidence": "low", "assessed_on": "2026-09-09", "onet_soc": "53-3032.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/truck-driver" }, { "slug": "graphic-designer", "name": "Graphic designer", "sector": "creative", "impact_index": 71, "impact_index_confidence": "medium", "assessed_on": "2026-09-09", "onet_soc": "27-1024.00 (exact)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/graphic-designer" }, { "slug": "junior-software-developer", "name": "Junior software developer", "sector": "tech", "impact_index": 62, "impact_index_confidence": "medium", "assessed_on": "2026-09-09", "onet_soc": "15-1252.00 (broader)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/junior-software-developer" }, { "slug": "paralegal", "name": "Paralegal", "sector": "public", "impact_index": 70, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "23-2011.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/paralegal" }, { "slug": "lawyer", "name": "Lawyer", "sector": "public", "impact_index": 48, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "23-1011.00 (exact)", "tasks": 5, "evidenced_tasks": 4, "url": "https://flyvolo.ai/en/careers/lawyer" }, { "slug": "journalist", "name": "Journalist", "sector": "creative", "impact_index": 58, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "27-3023.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/journalist" }, { "slug": "copywriter", "name": "Copywriter", "sector": "creative", "impact_index": 76, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "27-3043.00 (broader)", "tasks": 5, "evidenced_tasks": 1, "url": "https://flyvolo.ai/en/careers/copywriter" }, { "slug": "marketing-specialist", "name": "Marketing specialist", "sector": "business", "impact_index": 66, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "13-1161.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/marketing-specialist" }, { "slug": "data-analyst", "name": "Data analyst", "sector": "tech", "impact_index": 63, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "15-2051.01 (closest); 15-2051.00 (closest)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/data-analyst" }, { "slug": "financial-analyst", "name": "Financial analyst", "sector": "business", "impact_index": 60, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "13-2051.00 (exact)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/financial-analyst" }, { "slug": "hr-recruiter", "name": "HR / recruiter", "sector": "business", "impact_index": 61, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "13-1071.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/hr-recruiter" }, { "slug": "school-teacher", "name": "School teacher", "sector": "public", "impact_index": 34, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "25-2031.00 (narrower); 25-2021.00 (narrower)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/school-teacher" }, { "slug": "registered-nurse", "name": "Registered nurse", "sector": "health", "impact_index": 22, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "29-1141.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/registered-nurse" }, { "slug": "pharmacist", "name": "Pharmacist", "sector": "health", "impact_index": 44, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "29-1051.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/pharmacist" }, { "slug": "architect", "name": "Architect", "sector": "trades", "impact_index": 41, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "17-1011.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/architect" }, { "slug": "product-designer", "name": "Product / UX designer", "sector": "tech", "impact_index": 52, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "15-1255.00 (closest)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/product-designer" }, { "slug": "video-editor", "name": "Video editor", "sector": "creative", "impact_index": 69, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "27-4032.00 (exact)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/video-editor" }, { "slug": "software-tester", "name": "Software tester / QA engineer", "sector": "tech", "impact_index": 72, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "15-1253.00 (exact)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/software-tester" }, { "slug": "administrative-assistant", "name": "Administrative assistant", "sector": "business", "impact_index": 73, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "43-6014.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/administrative-assistant" }, { "slug": "bank-teller", "name": "Bank teller", "sector": "retail", "impact_index": 77, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "43-3071.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/bank-teller" }, { "slug": "retail-cashier", "name": "Retail cashier / shop assistant", "sector": "retail", "impact_index": 70, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "41-2011.00 (exact)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/retail-cashier" }, { "slug": "warehouse-worker", "name": "Warehouse worker", "sector": "transport", "impact_index": 56, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "53-7065.00 (narrower); 53-7062.00 (narrower)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/warehouse-worker" }, { "slug": "assembly-line-worker", "name": "Assembly line worker", "sector": "trades", "impact_index": 50, "impact_index_confidence": "low", "assessed_on": "2026-09-10", "onet_soc": "51-2092.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/assembly-line-worker" }, { "slug": "electrician", "name": "Electrician", "sector": "trades", "impact_index": 18, "impact_index_confidence": "medium", "assessed_on": "2026-09-10", "onet_soc": "47-2111.00 (exact)", "tasks": 5, "evidenced_tasks": 1, "url": "https://flyvolo.ai/en/careers/electrician" }, { "slug": "chef", "name": "Chef / cook", "sector": "retail", "impact_index": 33, "impact_index_confidence": "low", "assessed_on": "2026-09-10", "onet_soc": "35-1011.00 (narrower); 35-2014.00 (narrower)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/chef" }, { "slug": "delivery-rider", "name": "Delivery rider / courier", "sector": "transport", "impact_index": 40, "impact_index_confidence": "low", "assessed_on": "2026-09-10", "onet_soc": "43-5021.00 (closest)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/delivery-rider" }, { "slug": "ride-hail-driver", "name": "Ride-hail / taxi driver", "sector": "transport", "impact_index": 47, "impact_index_confidence": "low", "assessed_on": "2026-09-10", "onet_soc": "53-3054.00 (closest); 53-3053.00 (closest)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/ride-hail-driver" }, { "slug": "sales-account-manager", "name": "Sales / account manager", "sector": "business", "impact_index": 58, "impact_index_confidence": "medium", "assessed_on": "2026-09-11", "onet_soc": "41-4012.00 (closest); 41-3091.00 (closest)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/sales-account-manager" }, { "slug": "first-line-manager", "name": "First-line manager / team supervisor", "sector": "business", "impact_index": 41, "impact_index_confidence": "low", "assessed_on": "2026-09-11", "onet_soc": "43-1011.00 (closest)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/first-line-manager" }, { "slug": "operations-coordinator", "name": "Operations coordinator", "sector": "business", "impact_index": 62, "impact_index_confidence": "low", "assessed_on": "2026-09-11", "onet_soc": "43-5061.00 (closest); 13-1081.00 (closest)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/operations-coordinator" }, { "slug": "ai-implementation-lead", "name": "AI implementation lead", "sector": "tech", "impact_index": 16, "impact_index_confidence": "low", "assessed_on": "2026-09-11", "onet_soc": "none", "tasks": 6, "evidenced_tasks": 4, "url": "https://flyvolo.ai/en/careers/ai-implementation-lead" }, { "slug": "business-systems-owner", "name": "Business systems owner", "sector": "tech", "impact_index": 52, "impact_index_confidence": "low", "assessed_on": "2026-09-11", "onet_soc": "15-1211.00 (closest)", "tasks": 5, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/business-systems-owner" }, { "slug": "partnerships-manager", "name": "Partnerships / channel manager", "sector": "business", "impact_index": 44, "impact_index_confidence": "low", "assessed_on": "2026-09-11", "onet_soc": "11-2022.00 (closest)", "tasks": 6, "evidenced_tasks": 1, "url": "https://flyvolo.ai/en/careers/partnerships-manager" }, { "slug": "experienced-software-engineer", "name": "Experienced software engineer", "sector": "tech", "impact_index": 46, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "15-1252.00 (broader)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/experienced-software-engineer" }, { "slug": "frontend-developer", "name": "Frontend developer", "sector": "tech", "impact_index": 66, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "15-1254.00 (closest)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/frontend-developer" }, { "slug": "backend-developer", "name": "Backend developer", "sector": "tech", "impact_index": 54, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "15-1252.00 (broader)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/backend-developer" }, { "slug": "data-engineer", "name": "Data engineer", "sector": "tech", "impact_index": 50, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "15-1243.01 (closest); 15-1243.00 (closest)", "tasks": 6, "evidenced_tasks": 4, "url": "https://flyvolo.ai/en/careers/data-engineer" }, { "slug": "product-manager", "name": "Product manager", "sector": "tech", "impact_index": 45, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "none", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/product-manager" }, { "slug": "machine-learning-engineer", "name": "Machine learning engineer", "sector": "tech", "impact_index": 47, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "15-2051.00 (closest); 15-1221.00 (closest)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/machine-learning-engineer" }, { "slug": "radiologist", "name": "Radiologist", "sector": "health", "impact_index": 51, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "29-1224.00 (exact)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/radiologist" }, { "slug": "real-estate-agent", "name": "Real estate agent", "sector": "retail", "impact_index": 58, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "41-9022.00 (exact)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/real-estate-agent" }, { "slug": "auditor", "name": "Auditor", "sector": "business", "impact_index": 56, "impact_index_confidence": "low", "assessed_on": "2026-09-12", "onet_soc": "13-2011.00 (broader)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/auditor" }, { "slug": "insurance-claims-handler", "name": "Insurance claims handler", "sector": "business", "impact_index": 63, "impact_index_confidence": "low", "assessed_on": "2026-09-13", "onet_soc": "13-1031.00 (exact)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/insurance-claims-handler" }, { "slug": "counsellor", "name": "Counsellor / therapist", "sector": "health", "impact_index": 41, "impact_index_confidence": "low", "assessed_on": "2026-09-13", "onet_soc": "21-1014.00 (exact)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/counsellor" }, { "slug": "government-service-clerk", "name": "Government service clerk", "sector": "public", "impact_index": 66, "impact_index_confidence": "low", "assessed_on": "2026-09-13", "onet_soc": "43-4061.00 (closest)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/government-service-clerk" }, { "slug": "procurement-specialist", "name": "Procurement / supply chain specialist", "sector": "business", "impact_index": 59, "impact_index_confidence": "low", "assessed_on": "2026-09-13", "onet_soc": "13-1023.00 (exact)", "tasks": 6, "evidenced_tasks": 1, "url": "https://flyvolo.ai/en/careers/procurement-specialist" }, { "slug": "ai-researcher", "name": "AI researcher", "sector": "tech", "impact_index": 54, "impact_index_confidence": "low", "assessed_on": "2026-09-13", "onet_soc": "15-1221.00 (closest)", "tasks": 6, "evidenced_tasks": 5, "url": "https://flyvolo.ai/en/careers/ai-researcher" }, { "slug": "general-practitioner", "name": "General practitioner / primary care doctor", "sector": "health", "impact_index": 34, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "29-1215.00 (exact)", "tasks": 5, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/general-practitioner" }, { "slug": "care-worker", "name": "Care worker / nursing assistant", "sector": "health", "impact_index": 17, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "31-1131.00 (narrower); 31-1122.00 (narrower)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/care-worker" }, { "slug": "medical-assistant", "name": "Medical assistant / clinic assistant", "sector": "health", "impact_index": 48, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "31-9092.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/medical-assistant" }, { "slug": "lab-technician", "name": "Medical laboratory technician", "sector": "health", "impact_index": 49, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "29-2012.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/lab-technician" }, { "slug": "physical-therapist", "name": "Physiotherapist / rehabilitation therapist", "sector": "health", "impact_index": 27, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "29-1123.00 (exact)", "tasks": 4, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/physical-therapist" }, { "slug": "retail-salesperson", "name": "Retail salesperson / shop assistant", "sector": "retail", "impact_index": 41, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "41-2031.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/retail-salesperson" }, { "slug": "waiter", "name": "Waiter / restaurant server", "sector": "retail", "impact_index": 33, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "35-3031.00 (exact)", "tasks": 4, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/waiter" }, { "slug": "cleaner", "name": "Cleaner / janitor", "sector": "trades", "impact_index": 29, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "37-2011.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/cleaner" }, { "slug": "security-guard", "name": "Security guard", "sector": "trades", "impact_index": 38, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "33-9032.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/security-guard" }, { "slug": "it-support-specialist", "name": "IT support specialist / helpdesk", "sector": "tech", "impact_index": 57, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "15-1232.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/it-support-specialist" }, { "slug": "cybersecurity-analyst", "name": "Security analyst (SOC)", "sector": "tech", "impact_index": 44, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "15-1212.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/cybersecurity-analyst" }, { "slug": "devops-engineer", "name": "DevOps / platform / SRE engineer", "sector": "tech", "impact_index": 46, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "15-1244.00 (closest)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/devops-engineer" }, { "slug": "technical-writer", "name": "Technical writer / documentation engineer", "sector": "tech", "impact_index": 63, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "27-3042.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/technical-writer" }, { "slug": "receptionist", "name": "Receptionist / front desk", "sector": "retail", "impact_index": 58, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "43-4171.00 (exact)", "tasks": 4, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/receptionist" }, { "slug": "management-consultant", "name": "Management consultant", "sector": "business", "impact_index": 55, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "13-1111.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/management-consultant" }, { "slug": "loan-officer", "name": "Loan officer / credit officer", "sector": "business", "impact_index": 61, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "13-2072.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/loan-officer" }, { "slug": "compliance-officer", "name": "Compliance officer", "sector": "public", "impact_index": 43, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "13-1041.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/compliance-officer" }, { "slug": "construction-worker", "name": "Construction worker", "sector": "trades", "impact_index": 24, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "47-2061.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/construction-worker" }, { "slug": "auto-mechanic", "name": "Auto mechanic / vehicle technician", "sector": "trades", "impact_index": 31, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "49-3023.00 (exact)", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/auto-mechanic" }, { "slug": "ecommerce-operator", "name": "E-commerce operations specialist", "sector": "business", "impact_index": 52, "impact_index_confidence": "low", "assessed_on": "2026-09-14", "onet_soc": "none", "tasks": 4, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/ecommerce-operator" }, { "slug": "train-driver", "name": "Metro train driver", "sector": "transport", "impact_index": 62, "impact_index_confidence": "medium", "assessed_on": "2026-09-21", "onet_soc": "53-4041.00 (exact)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/train-driver" }, { "slug": "port-worker", "name": "Container port worker", "sector": "transport", "impact_index": 58, "impact_index_confidence": "medium", "assessed_on": "2026-09-21", "onet_soc": "53-7021.00 (narrower); 53-7062.00 (narrower)", "tasks": 5, "evidenced_tasks": 4, "url": "https://flyvolo.ai/en/careers/port-worker" }, { "slug": "content-moderator", "name": "Content moderator", "sector": "creative", "impact_index": 76, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "none", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/content-moderator" }, { "slug": "air-traffic-controller", "name": "Air traffic controller", "sector": "transport", "impact_index": 29, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "53-2021.00 (exact)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/air-traffic-controller" }, { "slug": "airline-pilot", "name": "Airline pilot", "sector": "transport", "impact_index": 34, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "53-2011.00 (exact)", "tasks": 6, "evidenced_tasks": 3, "url": "https://flyvolo.ai/en/careers/airline-pilot" }, { "slug": "bus-driver", "name": "Bus driver", "sector": "transport", "impact_index": 36, "impact_index_confidence": "low", "assessed_on": "2026-09-23", "onet_soc": "53-3052.00 (exact)", "tasks": 6, "evidenced_tasks": 4, "url": "https://flyvolo.ai/en/careers/bus-driver" }, { "slug": "actuary", "name": "Actuary", "sector": "business", "impact_index": 52, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "15-2011.00 (exact)", "tasks": 6, "evidenced_tasks": 4, "url": "https://flyvolo.ai/en/careers/actuary" }, { "slug": "civil-engineer", "name": "Civil / structural engineer", "sector": "trades", "impact_index": 40, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "17-2051.00 (exact)", "tasks": 6, "evidenced_tasks": 4, "url": "https://flyvolo.ai/en/careers/civil-engineer" }, { "slug": "firefighter", "name": "Firefighter", "sector": "public", "impact_index": 26, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "33-2011.00 (exact)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/firefighter" }, { "slug": "cabin-crew", "name": "Cabin crew", "sector": "transport", "impact_index": 22, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "53-2031.00 (exact)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/cabin-crew" }, { "slug": "police-officer", "name": "Police officer", "sector": "public", "impact_index": 30, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "33-3051.00 (exact)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/police-officer" }, { "slug": "farmer", "name": "Farmer", "sector": "trades", "impact_index": 28, "impact_index_confidence": "medium", "assessed_on": "2026-09-23", "onet_soc": "11-9013.00 (exact)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/farmer" }, { "slug": "social-worker", "name": "Social worker", "sector": "public", "impact_index": 24, "impact_index_confidence": "medium", "assessed_on": "2026-09-24", "onet_soc": "21-1021.00 (narrower)", "tasks": 6, "evidenced_tasks": 2, "url": "https://flyvolo.ai/en/careers/social-worker" }, { "slug": "dentist", "name": "Dentist", "sector": "health", "impact_index": 26, "impact_index_confidence": "medium", "assessed_on": "2026-09-24", "onet_soc": "29-1021.00 (exact)", "tasks": 6, "evidenced_tasks": 4, "url": "https://flyvolo.ai/en/careers/dentist" } ], "tasks": [ { "occupation_slug": "accountant", "task_id": "data-entry", "task": "Entering and coding transactions", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Does not mean the headcount disappears. In small practices the same person does entry and advisory; automating the entry half changes the job's shape before it changes the job count.", "evidence_ids": "ev-20250318-accountant-1; ev-20260812-accountant-3", "url": "https://flyvolo.ai/en/careers/accountant#task-data-entry" }, { "occupation_slug": "accountant", "task_id": "reconciliation", "task": "Reconciliation", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "The unmatched tail does not shrink in proportion to the matched volume. A practice that automates 95% of matching still needs someone who can chase the 5% — and that person needs to have seen the other 95% to recognise what is wrong.", "evidence_ids": "ev-20260812-accountant-3", "url": "https://flyvolo.ai/en/careers/accountant#task-reconciliation" }, { "occupation_slug": "accountant", "task_id": "compliance-filing", "task": "Statutory filing and tax compliance", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "The liability argument protects the sign-off, not the preparation hours behind it. Expect fewer hours per filing, not fewer filings needing a named person.", "evidence_ids": "ev-20250318-accountant-1", "url": "https://flyvolo.ai/en/careers/accountant#task-compliance-filing" }, { "occupation_slug": "accountant", "task_id": "advisory", "task": "Explaining the numbers to decision-makers", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Being hard to automate is not the same as being in demand. Advisory work is concentrated in the senior half of the profession; a junior whose entry work disappeared does not automatically arrive here.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/accountant#task-advisory" }, { "occupation_slug": "accountant", "task_id": "exception-judgement", "task": "Judging the ambiguous case", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "The volume of ambiguous cases is a fraction of total transactions. This task protects the necessity of the role, not the number of hours it takes to do.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/accountant#task-exception-judgement" }, { "occupation_slug": "accountant", "task_id": "automation-oversight", "task": "Supervising the automation itself", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "New work is not the same as new headcount, and this task is usually absorbed by people already there rather than hired for. It also requires the judgement built by doing the work that is disappearing.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/accountant#task-automation-oversight" }, { "occupation_slug": "customer-service-representative", "task_id": "faq-answering", "task": "Answering repeat questions", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Deflecting the easy half does not halve the work; it removes the easy half and leaves a queue where every remaining contact is harder than average. Handle time per contact goes up even as headcount falls.", "evidence_ids": "ev-20190101-customer-service-representative-5; ev-20240227-customer-service-representative-1; ev-20250508-customer-service-representative-2; ev-20260701-customer-service-representative-4; ev-20260812-customer-service-representative-3", "url": "https://flyvolo.ai/en/careers/customer-service-representative#task-faq-answering" }, { "occupation_slug": "customer-service-representative", "task_id": "triage", "task": "Routing and triage", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Misrouting is cheap only when a person is waiting at the other end to catch it. As more downstream steps also automate, the cost of a routing error stops being recoverable and starts being a customer who gives up.", "evidence_ids": "ev-20240227-customer-service-representative-1; ev-20260701-customer-service-representative-4; ev-20260812-customer-service-representative-3", "url": "https://flyvolo.ai/en/careers/customer-service-representative#task-triage" }, { "occupation_slug": "customer-service-representative", "task_id": "angry-customer", "task": "Handling a customer who is already angry", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This protects the role, not the headcount. If automation absorbs 70% of contacts, the remaining escalation work can be done by far fewer people.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/customer-service-representative#task-angry-customer" }, { "occupation_slug": "customer-service-representative", "task_id": "exception-authority", "task": "Deciding an exception", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "The constraint here is corporate liability policy, not capability — and policy is exactly the kind of thing that changes once the savings are demonstrated elsewhere. Treat this as a delay, not a moat.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/customer-service-representative#task-exception-authority" }, { "occupation_slug": "customer-service-representative", "task_id": "quality-oversight", "task": "Reviewing what the bot said", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "One reviewer can audit the output of many bots, which is precisely why this task does not replace the headcount that the bots displaced. It is a smaller, more senior job.", "evidence_ids": "ev-20250508-customer-service-representative-2", "url": "https://flyvolo.ai/en/careers/customer-service-representative#task-quality-oversight" }, { "occupation_slug": "translator", "task_id": "bulk-translation", "task": "First-draft translation of routine text", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Quality is highly uneven across language pairs and domains. Low-resource pairs and specialised terminology remain much weaker than the headline impression suggests.", "evidence_ids": "ev-20250428-translator-1; ev-20250701-translator-2; ev-20260120-translator-3; ev-20260428-translator-4", "url": "https://flyvolo.ai/en/careers/translator#task-bulk-translation" }, { "occupation_slug": "translator", "task_id": "post-editing", "task": "Post-editing machine output", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "It is real work, but it is usually priced per word at a fraction of translation rates. More of the industry's hours, less of its income.", "evidence_ids": "ev-20250701-translator-2; ev-20260428-translator-4", "url": "https://flyvolo.ai/en/careers/translator#task-post-editing" }, { "occupation_slug": "translator", "task_id": "high-stakes", "task": "Translation where being wrong is expensive", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Certification protects the signature, not the volume of work behind it. Expect the same documents to need a named certifier at a fraction of the hours, which is a pay structure change more than a job count change.", "evidence_ids": "ev-20190301-translator-5", "url": "https://flyvolo.ai/en/careers/translator#task-high-stakes" }, { "occupation_slug": "translator", "task_id": "cultural-adaptation", "task": "Deciding what to change, not just how to say it", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Transcreation is a small, high-end slice of a market whose bulk is volume translation. Being safe in this slice says nothing about whether there is room in it for the translators displaced from the rest.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/translator#task-cultural-adaptation" }, { "occupation_slug": "translator", "task_id": "live-interpreting", "task": "Interpreting in the room", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "The preference for an accountable human is strongest in exactly the settings that are rarest — courts, clinics, negotiations. Routine interpreting, which is most of the hours, has fewer defenders.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/translator#task-live-interpreting" }, { "occupation_slug": "truck-driver", "task_id": "highway-driving", "task": "Highway driving", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "mobility", "does_not_establish": "Automating the highway middle does not automate the trip. Hub-to-hub models still require a human for the first and last segments — which changes where drivers work far more than whether they work.", "evidence_ids": "ev-20250501-truck-driver-1; ev-20260609-truck-driver-2; ev-20260701-truck-driver-3", "url": "https://flyvolo.ai/en/careers/truck-driver#task-highway-driving" }, { "occupation_slug": "truck-driver", "task_id": "yard-and-dock", "task": "Yard manoeuvring and docking", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "mobility; physical", "does_not_establish": "Yards are private property with a single operator, which makes them the easiest place to change the rules — repaint the lanes, clear the pedestrians, and the problem gets much smaller. This is a harder task, not a permanently unreachable one.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/truck-driver#task-yard-and-dock" }, { "occupation_slug": "truck-driver", "task_id": "load-responsibility", "task": "Being responsible for the load", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "physical; cognitive", "does_not_establish": "Someone must legally accept the freight, but that someone does not have to have driven it there. Hub models already split the two, and where they do, the acceptance work moves to dock staff on an hourly wage.", "evidence_ids": "ev-20231121-truck-driver-4", "url": "https://flyvolo.ai/en/careers/truck-driver#task-load-responsibility" }, { "occupation_slug": "truck-driver", "task_id": "exception-handling", "task": "Handling the trip going wrong", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "mobility; cognitive", "does_not_establish": "Exception rate falls as the network is engineered around the vehicle: fixed lanes, known receivers, scheduled windows. The operators deploying today choose routes where exceptions are already rare.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/truck-driver#task-exception-handling" }, { "occupation_slug": "truck-driver", "task_id": "remote-supervision", "task": "Remote supervision of autonomous fleets", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "mobility; cognitive", "does_not_establish": "The ratio matters and is not yet settled publicly. One supervisor per many trucks is the commercial premise; one per two trucks would not change the industry's labour picture much.", "evidence_ids": "ev-20231121-truck-driver-4", "url": "https://flyvolo.ai/en/careers/truck-driver#task-remote-supervision" }, { "occupation_slug": "graphic-designer", "task_id": "asset-production", "task": "Producing routine assets", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "This was often the billable base that funded the interesting work, especially at agencies and for freelancers. Losing it changes the economics before it changes the craft.", "evidence_ids": "ev-20240528-graphic-designer-1; ev-20251201-graphic-designer-4; ev-20260122-graphic-designer-2", "url": "https://flyvolo.ai/en/careers/graphic-designer#task-asset-production" }, { "occupation_slug": "graphic-designer", "task_id": "concept-generation", "task": "Generating visual options", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "If options are free, clients ask for more of them and pay for fewer. The task does not disappear; its billable value does, which is a pricing problem rather than a capability one.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/graphic-designer#task-concept-generation" }, { "occupation_slug": "graphic-designer", "task_id": "brief-interrogation", "task": "Working out what the client actually needs", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is senior work that used to be paid for out of the production hours underneath it. Remove the production and the interrogation still happens, but there is no longer an obvious thing to invoice for it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/graphic-designer#task-brief-interrogation" }, { "occupation_slug": "graphic-designer", "task_id": "taste-and-defence", "task": "Judging quality and defending the call", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Judgement becoming more valuable does not mean more people get to exercise it. A team that needed six designers to produce may need two to judge — the survivors are better paid and fewer.", "evidence_ids": "ev-20260728-graphic-designer-3", "url": "https://flyvolo.ai/en/careers/graphic-designer#task-taste-and-defence" }, { "occupation_slug": "graphic-designer", "task_id": "system-ownership", "task": "Owning the system, not the artefact", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Design systems are per-organisation, not per-designer. One system owner serves everyone who produces, so this role scales with the number of brands, not with the volume of output.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/graphic-designer#task-system-ownership" }, { "occupation_slug": "junior-software-developer", "task_id": "boilerplate", "task": "Writing routine code", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Generating it is not the same as owning it. Someone still has to decide it is correct, and that someone has historically learned to judge by writing it first.", "evidence_ids": "ev-20241029-junior-software-developer-1; ev-20260812-junior-software-developer-3", "url": "https://flyvolo.ai/en/careers/junior-software-developer#task-boilerplate" }, { "occupation_slug": "junior-software-developer", "task_id": "debugging", "task": "Debugging unfamiliar systems", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is where the ceiling on generated code shows up most clearly, and it is the skill that separates a junior from someone who can be left alone.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/junior-software-developer#task-debugging" }, { "occupation_slug": "junior-software-developer", "task_id": "review-generated", "task": "Reviewing generated code", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Reviewing well requires having written the thing being reviewed. If the writing work that trained that judgement is the work being automated, this task has a supply problem within a few years.", "evidence_ids": "ev-20241029-junior-software-developer-1", "url": "https://flyvolo.ai/en/careers/junior-software-developer#task-review-generated" }, { "occupation_slug": "junior-software-developer", "task_id": "requirements", "task": "Turning a vague request into a spec", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This is normally a senior responsibility; listing it as a junior task describes where the role is heading, not where most junior jobs are today. The intermediate rungs are what disappeared.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/junior-software-developer#task-requirements" }, { "occupation_slug": "junior-software-developer", "task_id": "system-design", "task": "Deciding how pieces fit together", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Saying seniors are safe is not a statement about entry. The path to senior ran through two years of routine code, and nothing has replaced that path.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/junior-software-developer#task-system-design" }, { "occupation_slug": "paralegal", "task_id": "document-review", "task": "First-pass document review", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Privilege calls and anything that will be sworn to still route through a human, because the professional liability does not transfer to a vendor.", "evidence_ids": "ev-20260310-paralegal-2; ev-20260801-paralegal-3", "url": "https://flyvolo.ai/en/careers/paralegal#task-document-review" }, { "occupation_slug": "paralegal", "task_id": "drafting-from-precedent", "task": "Drafting from precedent", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The context that makes the tool useful — which precedent, what this client agreed before — is firm knowledge that someone has to keep maintaining. Firms that automate the drafting without funding that maintenance get worse drafts faster.", "evidence_ids": "ev-20260310-paralegal-2", "url": "https://flyvolo.ai/en/careers/paralegal#task-drafting-from-precedent" }, { "occupation_slug": "paralegal", "task_id": "matter-management", "task": "Running the matter", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Operational knowledge lives in people until someone writes it down — and the pressure to write it down is exactly what automation creates. This task is protected by an undocumented state that will not stay undocumented.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/paralegal#task-matter-management" }, { "occupation_slug": "paralegal", "task_id": "client-and-witness-contact", "task": "Client and witness contact", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Confidentiality duties constrain which interface may be used, not whether a person is needed. Firms are already procuring tools that satisfy those duties, and when they do, the argument moves.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/paralegal#task-client-and-witness-contact" }, { "occupation_slug": "paralegal", "task_id": "verification-of-machine-output", "task": "Checking what the tools produced", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Whether this task is paid as skilled work or absorbed as unpaid diligence depends on the firm, and that decides whether it is an opportunity or just more load.", "evidence_ids": "ev-20230622-paralegal-1", "url": "https://flyvolo.ai/en/careers/paralegal#task-verification-of-machine-output" }, { "occupation_slug": "lawyer", "task_id": "legal-research", "task": "Legal research", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Research hours were how firms trained and billed juniors. Compressing them protects the partner's judgement while removing the apprenticeship that produced the next partner.", "evidence_ids": "ev-20260528-lawyer-2", "url": "https://flyvolo.ai/en/careers/lawyer#task-legal-research" }, { "occupation_slug": "lawyer", "task_id": "drafting-and-negotiation", "task": "Drafting and negotiating documents", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Negotiation is a small share of the hours on most matters; drafting is a large one. Being protected on the former does not protect the leverage model that made the matter profitable.", "evidence_ids": "ev-20250224-lawyer-3", "url": "https://flyvolo.ai/en/careers/lawyer#task-drafting-and-negotiation" }, { "occupation_slug": "lawyer", "task_id": "advice-and-judgement", "task": "Giving advice someone will act on", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A licensing rule is a policy choice, and policy choices are argued about when the cost difference becomes large enough. Several jurisdictions are already consulting on what non-lawyers may do.", "evidence_ids": "ev-20260821-lawyer-4", "url": "https://flyvolo.ai/en/careers/lawyer#task-advice-and-judgement" }, { "occupation_slug": "lawyer", "task_id": "advocacy", "task": "Advocacy and appearances", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Appearances are a small fraction of most lawyers' time and are concentrated in litigation. For the transactional majority this task offers no protection at all.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/lawyer#task-advocacy" }, { "occupation_slug": "lawyer", "task_id": "supervising-machine-work", "task": "Supervising machine-assisted work", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "This is an added duty, not an added role — it arrives as unbilled responsibility on people who already had a full week. Nobody is hiring a supervisor of machine work; they are extending the existing duty of competence.", "evidence_ids": "ev-20230622-lawyer-1", "url": "https://flyvolo.ai/en/careers/lawyer#task-supervising-machine-work" }, { "occupation_slug": "journalist", "task_id": "commodity-rewrites", "task": "Rewriting press releases and results", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This was often what junior reporters did while learning the beat. Removing it removes a training ground, not just a cost.", "evidence_ids": "ev-20250518-journalist-1; ev-20250722-journalist-2", "url": "https://flyvolo.ai/en/careers/journalist#task-commodity-rewrites" }, { "occupation_slug": "journalist", "task_id": "sourcing", "task": "Getting people to tell you things", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Relationships protect the reporter who already has them. They do nothing for the entry-level jobs where those relationships used to be built, and those are the jobs the industry cut first.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/journalist#task-sourcing" }, { "occupation_slug": "journalist", "task_id": "verification", "task": "Verifying claims and material", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "More verification work is not more verification jobs. It typically lands on the same reporters as an added step, and newsrooms that cannot fund a verification desk simply publish with less checking.", "evidence_ids": "ev-20250518-journalist-1", "url": "https://flyvolo.ai/en/careers/journalist#task-verification" }, { "occupation_slug": "journalist", "task_id": "drafting-and-structure", "task": "Writing the piece", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Editorial judgement sits with a shrinking number of senior people. Assistance speeds the writing of a piece without changing how many pieces an outlet can fund.", "evidence_ids": "ev-20260522-journalist-3", "url": "https://flyvolo.ai/en/careers/journalist#task-drafting-and-structure" }, { "occupation_slug": "journalist", "task_id": "synthetic-media-forensics", "task": "Detecting synthetic and manipulated media", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Specialist desks exist at large outlets and essentially nowhere else. For most journalists this is a skill that makes them more employable, not a job that exists to be applied for.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/journalist#task-synthetic-media-forensics" }, { "occupation_slug": "copywriter", "task_id": "volume-copy", "task": "Producing volume copy", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Cheap copy has flooded every channel, which has started to lower its effectiveness. That does not bring the old prices back, but it does create demand for someone who can tell why a piece is not working.", "evidence_ids": "ev-20240528-copywriter-1; ev-20260626-copywriter-3; ev-20260731-copywriter-2", "url": "https://flyvolo.ai/en/careers/copywriter#task-volume-copy" }, { "occupation_slug": "copywriter", "task_id": "positioning-and-message", "task": "Finding the message", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Message work is bought in small, senior quantities — often by the founder or a head of marketing rather than from a writer. The volume work that funded a copywriting career is the part that went.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/copywriter#task-positioning-and-message" }, { "occupation_slug": "copywriter", "task_id": "brand-voice-system", "task": "Defining and policing a voice", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "One voice owner per brand is the ceiling. This role is real and it is where experienced copywriters land, but there is exactly one of it per company.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/copywriter#task-brand-voice-system" }, { "occupation_slug": "copywriter", "task_id": "performance-iteration", "task": "Reading results and iterating", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Insight is scarce, but reading results is increasingly packaged into the ad platforms themselves. The task is protected by the analyst's context, not by any barrier the platforms cannot cross.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/copywriter#task-performance-iteration" }, { "occupation_slug": "copywriter", "task_id": "high-stakes-copy", "task": "Copy where being wrong costs money", "direction": "human_led", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive", "does_not_establish": "It is explicitly the smallest slice of the role — peripheral by weight. Being safe on the pricing page does not pay for a week.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/copywriter#task-high-stakes-copy" }, { "occupation_slug": "marketing-specialist", "task_id": "content-and-creative-production", "task": "Producing campaign content", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "When the platform generates the creative as part of placing the ad, the saving accrues to the platform's margin and the advertiser's budget, not to the marketer's time. The hours vanish without anything replacing what they paid for.", "evidence_ids": "ev-20240528-marketing-specialist-1; ev-20260801-marketing-specialist-3", "url": "https://flyvolo.ai/en/careers/marketing-specialist#task-content-and-creative-production" }, { "occupation_slug": "marketing-specialist", "task_id": "paid-media-operation", "task": "Running paid media", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Platform automation optimises for the platform's metrics. Knowing when the reported result is not the real business result is still a human job, and a well-paid one.", "evidence_ids": "ev-20260226-marketing-specialist-2", "url": "https://flyvolo.ai/en/careers/marketing-specialist#task-paid-media-operation" }, { "occupation_slug": "marketing-specialist", "task_id": "audience-understanding", "task": "Understanding who actually buys and why", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Knowing the customer is undisputed and rarely resourced. It is the first thing cut when a team shrinks, precisely because nothing breaks visibly that quarter.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/marketing-specialist#task-audience-understanding" }, { "occupation_slug": "marketing-specialist", "task_id": "strategy-and-allocation", "task": "Deciding where the money goes", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Allocation is one person's job on most teams. Protecting it protects the head of marketing, not the specialists whose production work was the rest of the department.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/marketing-specialist#task-strategy-and-allocation" }, { "occupation_slug": "marketing-specialist", "task_id": "agent-and-automation-operation", "task": "Operating the marketing automation itself", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "The site's own phrasing is that the production hours reappear 'partly' as supervision. Partly is the operative word: supervision of many automated channels is fewer hours than producing for them was.", "evidence_ids": "ev-20260226-marketing-specialist-2", "url": "https://flyvolo.ai/en/careers/marketing-specialist#task-agent-and-automation-operation" }, { "occupation_slug": "data-analyst", "task_id": "query-writing", "task": "Writing queries and building dashboards", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "'When the data is clean and the question is clear' is a large condition. In most companies neither holds, and the tools produce confident wrong answers on ambiguous schemas.", "evidence_ids": "ev-20240919-data-analyst-1; ev-20260720-data-analyst-2; ev-20260910-data-analyst-3", "url": "https://flyvolo.ai/en/careers/data-analyst#task-query-writing" }, { "occupation_slug": "data-analyst", "task_id": "question-framing", "task": "Working out what is really being asked", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Self-service tools let managers ask their own questions badly rather than ask an analyst. Framing stays valuable, but the analyst has to be in the room to supply it, and fewer are.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/data-analyst#task-question-framing" }, { "occupation_slug": "data-analyst", "task_id": "data-trust", "task": "Knowing when the data is lying", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Institutional memory of the data is built by working with it daily. If the daily querying moves to self-service, the memory that makes this task possible stops accumulating.", "evidence_ids": "ev-20260720-data-analyst-2; ev-20260910-data-analyst-3", "url": "https://flyvolo.ai/en/careers/data-analyst#task-data-trust" }, { "occupation_slug": "data-analyst", "task_id": "communicating-results", "task": "Making a number change a decision", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Credibility is personal and accrues slowly, which protects established analysts and no one else. A first analyst hire has none of it and now has fewer routine tasks to earn it with.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/data-analyst#task-communicating-results" }, { "occupation_slug": "data-analyst", "task_id": "semantic-layer-ownership", "task": "Owning the definitions the tools rely on", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Curating definitions is a smaller job than answering questions was, and it is one per organisation. Headcount being 'redirected' into it is a euphemism for a smaller team doing different work.", "evidence_ids": "ev-20260910-data-analyst-3", "url": "https://flyvolo.ai/en/careers/data-analyst#task-semantic-layer-ownership" }, { "occupation_slug": "financial-analyst", "task_id": "model-building", "task": "Building and updating models", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Structure design is the interesting hour in a week of updating. Automating the update removes the hours that justified the headcount, while leaving the hour that justifies the title.", "evidence_ids": "ev-20250520-financial-analyst-3; ev-20260910-financial-analyst-2", "url": "https://flyvolo.ai/en/careers/financial-analyst#task-model-building" }, { "occupation_slug": "financial-analyst", "task_id": "data-gathering-and-summaries", "task": "Gathering data and summarising filings", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "The summaries are only as good as the reader's ability to notice what was left out. Juniors who never read a full filing lose the instinct for what a summary hides.", "evidence_ids": "ev-20250623-financial-analyst-1; ev-20260910-financial-analyst-2", "url": "https://flyvolo.ai/en/careers/financial-analyst#task-data-gathering-and-summaries" }, { "occupation_slug": "financial-analyst", "task_id": "judgement-on-assumptions", "task": "Deciding what to assume", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Assumption-setting is the senior half of the job by definition. Analysts are not hired into it; they arrive after years of the modelling work that is being compressed.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/financial-analyst#task-judgement-on-assumptions" }, { "occupation_slug": "financial-analyst", "task_id": "stakeholder-communication", "task": "Explaining to the decision-maker", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "The meeting in which capital is committed happens a few times a quarter. It protects the seat at the table, not the work that filled the rest of the calendar.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/financial-analyst#task-stakeholder-communication" }, { "occupation_slug": "financial-analyst", "task_id": "tool-orchestration", "task": "Orchestrating the analytical tooling", "direction": "emerging", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive; process", "does_not_establish": "Explicitly peripheral by weight, and typically absorbed by one person on a team as an extra. It is not a job posting.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/financial-analyst#task-tool-orchestration" }, { "occupation_slug": "hr-recruiter", "task_id": "sourcing-and-screening", "task": "Sourcing and screening candidates", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Automated screening is regulated in several jurisdictions as high-risk decision-making, and discrimination liability stays with the employer. Adoption is constrained by law and by lawsuits, not just by capability.", "evidence_ids": "ev-20240801-hr-recruiter-1; ev-20260122-hr-recruiter-5; ev-20260701-hr-recruiter-4", "url": "https://flyvolo.ai/en/careers/hr-recruiter#task-sourcing-and-screening" }, { "occupation_slug": "hr-recruiter", "task_id": "hr-administration", "task": "HR administration", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "The exceptions that remain are a small share of volume and require the most experienced people. Departments shrink at the bottom and keep the top, which is a worse ladder rather than a stable one.", "evidence_ids": "ev-20250501-hr-recruiter-3", "url": "https://flyvolo.ai/en/careers/hr-recruiter#task-hr-administration" }, { "occupation_slug": "hr-recruiter", "task_id": "assessment-and-decision", "task": "Deciding whom to hire", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The decision stays human while everything feeding it — sourcing, screening, scheduling, note-taking — does not. Protecting the final call protects one hour in a process that used to take twenty.", "evidence_ids": "ev-20260122-hr-recruiter-5; ev-20260514-hr-recruiter-2", "url": "https://flyvolo.ai/en/careers/hr-recruiter#task-assessment-and-decision" }, { "occupation_slug": "hr-recruiter", "task_id": "employee-relations", "task": "Handling difficult situations", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Demand being stable is not the same as demand growing. This is a specialist function inside HR; the recruiting and administration roles that made up the headcount are elsewhere.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/hr-recruiter#task-employee-relations" }, { "occupation_slug": "hr-recruiter", "task_id": "workforce-redesign", "task": "Redesigning jobs around automation", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "It is project work that peaks during a transition and subsides. Building a career on helping other people through redundancy has an obvious ceiling.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/hr-recruiter#task-workforce-redesign" }, { "occupation_slug": "school-teacher", "task_id": "lesson-preparation", "task": "Preparing lessons and materials", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Time returned to a teacher is rarely time removed from a school's staffing. It shows up as capacity for the duties that were already being squeezed, not as fewer teachers.", "evidence_ids": "ev-20250513-school-teacher-3; ev-20260309-school-teacher-1; ev-20260402-school-teacher-4", "url": "https://flyvolo.ai/en/careers/school-teacher#task-lesson-preparation" }, { "occupation_slug": "school-teacher", "task_id": "marking-and-feedback", "task": "Marking and written feedback", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Automated feedback on writing is only as good as the rubric, and students learn quickly to write for the machine. Schools that let it replace rather than assist the teacher's reading have seen the quality of feedback fall.", "evidence_ids": "ev-20260402-school-teacher-4", "url": "https://flyvolo.ai/en/careers/school-teacher#task-marking-and-feedback" }, { "occupation_slug": "school-teacher", "task_id": "classroom-management", "task": "Running the room", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive; physical", "does_not_establish": "This protects the profession's headcount far better than its working conditions. The tasks being automated are the ones teachers do at home; the ones that exhaust them are untouched.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/school-teacher#task-classroom-management" }, { "occupation_slug": "school-teacher", "task_id": "explaining-and-adapting-live", "task": "Explaining, and noticing who did not get it", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Tutoring systems work best for motivated students with access — which is a statement about who benefits, not about whether classroom teaching changes. The gap they widen is between students, not between teachers and machines.", "evidence_ids": "ev-20260724-school-teacher-2", "url": "https://flyvolo.ai/en/careers/school-teacher#task-explaining-and-adapting-live" }, { "occupation_slug": "school-teacher", "task_id": "orchestrating-learning-tools", "task": "Orchestrating tutoring tools", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "cognitive", "does_not_establish": "This arrives as new curriculum with no extra time, which the site's own reasoning says explicitly. New responsibility without new hours is a workload change, not a new role.", "evidence_ids": "ev-20250513-school-teacher-3; ev-20260309-school-teacher-1; ev-20260724-school-teacher-2", "url": "https://flyvolo.ai/en/careers/school-teacher#task-orchestrating-learning-tools" }, { "occupation_slug": "registered-nurse", "task_id": "documentation", "task": "Documentation and handover notes", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Time returned on a short-staffed ward becomes more patients per nurse, not a shorter shift. Whether the hour is a gain for the nurse or for the roster is a funding decision, not a technical one.", "evidence_ids": "ev-20251104-registered-nurse-1; ev-20260508-registered-nurse-2", "url": "https://flyvolo.ai/en/careers/registered-nurse#task-documentation" }, { "occupation_slug": "registered-nurse", "task_id": "monitoring-and-escalation", "task": "Monitoring and knowing when to escalate", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Alarms that are wrong often enough to need judgement are also wrong often enough to cause alarm fatigue. More tools on the ward can degrade this task rather than support it.", "evidence_ids": "ev-20210621-registered-nurse-4", "url": "https://flyvolo.ai/en/careers/registered-nurse#task-monitoring-and-escalation" }, { "occupation_slug": "registered-nurse", "task_id": "hands-on-care", "task": "Hands-on care", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical", "does_not_establish": "Being physically irreplaceable has never protected nursing pay or staffing levels — those are set by health budgets. The constraint on this task is money, and money moves for reasons unrelated to capability.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/registered-nurse#task-hands-on-care" }, { "occupation_slug": "registered-nurse", "task_id": "patient-and-family-communication", "task": "Talking to patients and families", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The site's own reasoning says the constraint is staffing, which is a funding problem. A task nobody intends to automate can still be a task nobody is paid enough to do.", "evidence_ids": "ev-20251011-registered-nurse-3", "url": "https://flyvolo.ai/en/careers/registered-nurse#task-patient-and-family-communication" }, { "occupation_slug": "registered-nurse", "task_id": "supervising-clinical-tools", "task": "Working with clinical decision tools", "direction": "emerging", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive", "does_not_establish": "Peripheral by weight and formalised by hospitals as part of the existing role. It adds responsibility to a shift that is already full.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/registered-nurse#task-supervising-clinical-tools" }, { "occupation_slug": "pharmacist", "task_id": "dispensing", "task": "Dispensing and counting", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "physical; process", "does_not_establish": "Where it has not automated, the reason is capital cost at small pharmacies — and capital cost falls. Independent pharmacies are exposed on a delay, not exempt.", "evidence_ids": "ev-20250520-pharmacist-1", "url": "https://flyvolo.ai/en/careers/pharmacist#task-dispensing" }, { "occupation_slug": "pharmacist", "task_id": "prescription-verification", "task": "Checking the prescription is safe", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Regulated accountability protects the signature on the release, not the number of pharmacists a chain employs per thousand prescriptions. Central-fill models already concentrate that signature.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/pharmacist#task-prescription-verification" }, { "occupation_slug": "pharmacist", "task_id": "patient-counselling", "task": "Counselling the patient", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Counselling is valued and rarely paid for directly. In systems that reimburse dispensing rather than advice, this task is the first thing squeezed when the dispensing fee falls.", "evidence_ids": "ev-20260511-pharmacist-2", "url": "https://flyvolo.ai/en/careers/pharmacist#task-patient-counselling" }, { "occupation_slug": "pharmacist", "task_id": "clinical-services", "task": "Clinical services", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; physical", "does_not_establish": "This expansion is driven by policy and workforce shortage, both of which can reverse. It is a bet on health systems continuing to be short of doctors.", "evidence_ids": "ev-20250520-pharmacist-1; ev-20260511-pharmacist-2", "url": "https://flyvolo.ai/en/careers/pharmacist#task-clinical-services" }, { "occupation_slug": "pharmacist", "task_id": "supply-and-shortage-management", "task": "Managing supply and shortages", "direction": "human_led", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive; process", "does_not_establish": "Explicitly peripheral, and it exists because supply chains are broken. Building a case for the role on other people's dysfunction is fragile.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/pharmacist#task-supply-and-shortage-management" }, { "occupation_slug": "architect", "task_id": "visualisation", "task": "Renderings and presentation images", "direction": "automating", "basis": "evidenced", "weight": "peripheral", "technologies": "cognitive", "does_not_establish": "This was a specialist sub-trade with its own studios and its own people. Practices absorbing it in-house is good for practices and terminal for visualisers.", "evidence_ids": "ev-20260701-architect-3", "url": "https://flyvolo.ai/en/careers/architect#task-visualisation" }, { "occupation_slug": "architect", "task_id": "documentation-and-detailing", "task": "Construction documentation", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Documentation is where most of a practice's billable hours sit and where most architects spend their careers. Compressing it changes the economics of running a practice long before it changes who signs the drawings.", "evidence_ids": "ev-20231218-architect-1; ev-20260630-architect-2", "url": "https://flyvolo.ai/en/careers/architect#task-documentation-and-detailing" }, { "occupation_slug": "architect", "task_id": "brief-and-concept", "task": "Understanding the brief and finding the idea", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Concept work is what architects want to do and a small share of what they are paid for. Protection here does not fund a practice.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/architect#task-brief-and-concept" }, { "occupation_slug": "architect", "task_id": "approvals-and-coordination", "task": "Approvals and coordinating the team", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Negotiating with authorities is protected by how authorities work, which is a rule set, not a law of nature. Digital planning submission regimes are already narrowing what has to be argued in person.", "evidence_ids": "ev-20231218-architect-1", "url": "https://flyvolo.ai/en/careers/architect#task-approvals-and-coordination" }, { "occupation_slug": "architect", "task_id": "site-and-construction", "task": "Site inspection and construction-phase decisions", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive; physical", "does_not_establish": "Site work is a phase, not a job, and on many projects it is contracted out to a different role entirely. Being safe on site says little about the studio jobs that make up the profession.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/architect#task-site-and-construction" }, { "occupation_slug": "product-designer", "task_id": "screen-production", "task": "Producing screens and components", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Pixel production was the paid apprenticeship. Removing it does not just remove junior jobs; it removes how anyone becomes the senior designer the same team still needs.", "evidence_ids": "ev-20240702-product-designer-1", "url": "https://flyvolo.ai/en/careers/product-designer#task-screen-production" }, { "occupation_slug": "product-designer", "task_id": "problem-definition", "task": "Defining the right problem", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Defining the problem is usually shared with product management, and on small teams it is not the designer who wins that argument. Being valuable is not the same as being the one who decides.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/product-designer#task-problem-definition" }, { "occupation_slug": "product-designer", "task_id": "research-and-testing", "task": "Research and usability testing", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Research is the first budget cut in most downturns, regardless of whether synthetic users are any good. Its exposure is commercial before it is technical.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/product-designer#task-research-and-testing" }, { "occupation_slug": "product-designer", "task_id": "cross-functional-alignment", "task": "Getting the team to agree", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Alignment work grows with team size and disappears with the team. It protects the designer who is already trusted, not the role's headcount.", "evidence_ids": "ev-20260515-product-designer-2", "url": "https://flyvolo.ai/en/careers/product-designer#task-cross-functional-alignment" }, { "occupation_slug": "product-designer", "task_id": "designing-for-generative-interfaces", "task": "Designing for non-deterministic products", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Scarcity here is temporary by construction — almost no one has done it for long because it is new, and that gap closes as everyone gets the same few years of practice.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/product-designer#task-designing-for-generative-interfaces" }, { "occupation_slug": "video-editor", "task_id": "assembly-and-rough-cut", "task": "Logging footage and assembling a rough cut", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "This was the block of hours that junior editors were paid for while learning the judgement in the next task. Its disappearance is a training problem as much as an hours problem.", "evidence_ids": "ev-20240804-video-editor-3", "url": "https://flyvolo.ai/en/careers/video-editor#task-assembly-and-rough-cut" }, { "occupation_slug": "video-editor", "task_id": "formatting-and-versioning", "task": "Versions, formats and captions", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Freelancers priced by deliverable, and versioning multiplied deliverables. Losing it removes a volume business, which hurts most at the bottom of the market.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/video-editor#task-formatting-and-versioning" }, { "occupation_slug": "video-editor", "task_id": "story-and-pacing", "task": "Finding the story and the rhythm", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "The market for people who make this call well has not shrunk — but it was always small, and it is now the only door. Everyone displaced from assembly is trying to walk through it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/video-editor#task-story-and-pacing" }, { "occupation_slug": "video-editor", "task_id": "client-direction", "task": "Reading and managing the client", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Clients who can generate rough versions themselves also anchor on them, which makes this task larger and more frustrating without making it better paid.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/video-editor#task-client-direction" }, { "occupation_slug": "video-editor", "task_id": "generative-and-hybrid-production", "task": "Directing generated footage", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Disclosure rules, client policies and audience trust constrain where generated footage may be used, and the constraints differ sharply between advertising, corporate and journalistic work.", "evidence_ids": "ev-20241112-video-editor-1; ev-20260716-video-editor-2", "url": "https://flyvolo.ai/en/careers/video-editor#task-generative-and-hybrid-production" }, { "occupation_slug": "software-tester", "task_id": "manual-regression", "task": "Manual regression testing", "direction": "automating", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "This is explicitly the largest block of hours in most QA roles. Nothing in the remaining tasks is large enough to absorb the people whose week it filled.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/software-tester#task-manual-regression" }, { "occupation_slug": "software-tester", "task_id": "test-case-writing", "task": "Writing test cases and automation", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Generated tests confirm what the code does rather than what it should. Teams that measure coverage rather than defects will conclude this task is solved, which is a measurement problem that hurts testers.", "evidence_ids": "ev-20240214-software-tester-1", "url": "https://flyvolo.ai/en/careers/software-tester#task-test-case-writing" }, { "occupation_slug": "software-tester", "task_id": "exploratory-and-adversarial", "task": "Exploratory and adversarial testing", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The hypothesis about where to look is built by having done the regression work that is disappearing. This task is protected by experience that the pipeline no longer produces.", "evidence_ids": "ev-20260414-software-tester-2", "url": "https://flyvolo.ai/en/careers/software-tester#task-exploratory-and-adversarial" }, { "occupation_slug": "software-tester", "task_id": "release-judgement", "task": "Deciding whether it ships", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "On many teams this call belongs to an engineering manager, not a tester. Where it does, protecting it protects someone else's job.", "evidence_ids": "ev-20260414-software-tester-2", "url": "https://flyvolo.ai/en/careers/software-tester#task-release-judgement" }, { "occupation_slug": "software-tester", "task_id": "testing-model-based-systems", "task": "Testing systems with a model inside", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Few practitioners is a statement about now. The discipline is young enough that its eventual size is unknown, and it may end up inside the model teams rather than in QA.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/software-tester#task-testing-model-based-systems" }, { "occupation_slug": "administrative-assistant", "task_id": "scheduling-and-correspondence", "task": "Scheduling and correspondence", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "This was the defining task of the role, and it is being absorbed into software the executive already has open — which means the substitution needs no procurement decision and no budget line. That is the fastest kind of displacement.", "evidence_ids": "ev-20241231-administrative-assistant-1; ev-20260827-administrative-assistant-2", "url": "https://flyvolo.ai/en/careers/administrative-assistant#task-scheduling-and-correspondence" }, { "occupation_slug": "administrative-assistant", "task_id": "expenses-travel-documents", "task": "Expenses, travel and documents", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "Exceptions and disputes are a fraction of the volume and they arrive unpredictably. A role cannot be staffed on a task that is 10% of the hours but needs someone available all week.", "evidence_ids": "ev-20260413-administrative-assistant-3", "url": "https://flyvolo.ai/en/careers/administrative-assistant#task-expenses-travel-documents" }, { "occupation_slug": "administrative-assistant", "task_id": "gatekeeping-and-judgement", "task": "Gatekeeping and knowing what matters", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This distinguishes an executive assistant from an administrator — which is another way of saying it protects the senior tier and not the tier below it. Most people in this occupation are in the tier below.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/administrative-assistant#task-gatekeeping-and-judgement" }, { "occupation_slug": "administrative-assistant", "task_id": "events-and-logistics", "task": "Events and physical logistics", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "physical; cognitive", "does_not_establish": "The site's own reasoning notes this is what gets missed when organisations cut administrators — meaning organisations cut them anyway and absorb the loss. Being missed is not the same as being retained.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/administrative-assistant#task-events-and-logistics" }, { "occupation_slug": "administrative-assistant", "task_id": "operating-the-executives-tools", "task": "Operating the executive's assistants", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "cognitive; process", "does_not_establish": "Peripheral by weight and it lands on whoever already knew the preferences. It is a reason to keep one assistant, not a reason to keep the team.", "evidence_ids": "ev-20260827-administrative-assistant-2", "url": "https://flyvolo.ai/en/careers/administrative-assistant#task-operating-the-executives-tools" }, { "occupation_slug": "bank-teller", "task_id": "routine-transactions", "task": "Routine transactions", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "process", "does_not_establish": "The contraction has been gradual and uneven, and several banks have reversed branch closures after losing customers. Automation of transactions did not eliminate the role; it changed the mix of who comes to the counter and why.", "evidence_ids": "ev-20250130-bank-teller-1; ev-20260113-bank-teller-3", "url": "https://flyvolo.ai/en/careers/bank-teller#task-routine-transactions" }, { "occupation_slug": "bank-teller", "task_id": "identity-and-compliance", "task": "Identity, fraud and compliance checks", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Regulators expecting branches to play this role depends on branches existing. Where networks close, the expectation migrates to phone and app channels rather than protecting the counter.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/bank-teller#task-identity-and-compliance" }, { "occupation_slug": "bank-teller", "task_id": "helping-with-complexity", "task": "Helping with the complicated cases", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Exceptions are a rising share of what walks in the door because the routine left, not because there are more exceptions. A branch serving only exceptions serves far fewer people and needs far fewer staff.", "evidence_ids": "ev-20260222-bank-teller-2", "url": "https://flyvolo.ai/en/careers/bank-teller#task-helping-with-complexity" }, { "occupation_slug": "bank-teller", "task_id": "sales-and-referral", "task": "Recognising a need and referring", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Referral targets make the role measurable, which cuts both ways: a counter that does not hit them is easier to close. This task ties the job to a sales number rather than to a service need.", "evidence_ids": "ev-20260222-bank-teller-2", "url": "https://flyvolo.ai/en/careers/bank-teller#task-sales-and-referral" }, { "occupation_slug": "bank-teller", "task_id": "digital-coaching", "task": "Teaching customers the digital channels", "direction": "emerging", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive", "does_not_establish": "The task is explicitly about teaching customers to need the branch less. It is stable only for as long as there is a cohort that has not made the transition, and that cohort shrinks every year.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/bank-teller#task-digital-coaching" }, { "occupation_slug": "retail-cashier", "task_id": "checkout", "task": "Scanning and taking payment", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "process", "does_not_establish": "Several major retailers have pulled back self-checkout after theft and customer frustration outweighed the savings. Adoption has oscillated, which is a useful reminder that deployment is a business decision, not a technology inevitability.", "evidence_ids": "ev-20231110-retail-cashier-1; ev-20260224-retail-cashier-2", "url": "https://flyvolo.ai/en/careers/retail-cashier#task-checkout" }, { "occupation_slug": "retail-cashier", "task_id": "shelf-and-stock", "task": "Restocking and shelf work", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical", "does_not_establish": "Shelf work is physically hard, low-paid and scheduled around peaks. Being the least automatable task in the shop is not the same as being a job people can build a life on.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/retail-cashier#task-shelf-and-stock" }, { "occupation_slug": "retail-cashier", "task_id": "customer-help", "task": "Helping customers", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Retailers measuring the effect on satisfaction and reversing course is one set of retailers. Others measured the same thing and kept the savings, and the site has evidence of both.", "evidence_ids": "ev-20260224-retail-cashier-2", "url": "https://flyvolo.ai/en/careers/retail-cashier#task-customer-help" }, { "occupation_slug": "retail-cashier", "task_id": "loss-prevention", "task": "Loss prevention and the difficult customer", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; physical", "does_not_establish": "This task grew because self-checkout created the theft it now polices. It is work generated by an automation decision, and it is more dangerous and no better paid than the work it replaced.", "evidence_ids": "ev-20260729-retail-cashier-3", "url": "https://flyvolo.ai/en/careers/retail-cashier#task-loss-prevention" }, { "occupation_slug": "retail-cashier", "task_id": "order-fulfilment", "task": "Picking online orders in store", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "physical; process", "does_not_establish": "In-store picking exists because delivery economics have not been solved, not because the work is valuable. Dark stores and automated fulfilment centres are the direction of travel where volume justifies them.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/retail-cashier#task-order-fulfilment" }, { "occupation_slug": "warehouse-worker", "task_id": "goods-movement", "task": "Moving goods around the building", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "mobility; physical", "does_not_establish": "Goods-to-person robotics moved the walking, not the grasping, and it did so by rebuilding the building around the robot. The displacement follows capital investment, so it arrives suddenly per site rather than gradually across the industry.", "evidence_ids": "ev-20250701-warehouse-worker-1; ev-20260305-warehouse-worker-4; ev-20260701-warehouse-worker-3", "url": "https://flyvolo.ai/en/careers/warehouse-worker#task-goods-movement" }, { "occupation_slug": "warehouse-worker", "task_id": "picking-and-packing", "task": "Picking items and packing orders", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "Progress on grasping is real and fast. This is the task on the site most likely to have moved by the next assessment, and we will revise it when deployments, not demonstrations, change.", "evidence_ids": "ev-20260305-warehouse-worker-4; ev-20260604-warehouse-worker-2; ev-20260701-warehouse-worker-3", "url": "https://flyvolo.ai/en/careers/warehouse-worker#task-picking-and-packing" }, { "occupation_slug": "warehouse-worker", "task_id": "exception-handling", "task": "Handling what the system cannot", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "physical; cognitive", "does_not_establish": "The more automated the building, the more the remaining human work is exception handling — but also the fewer humans there are to do it. This task grows as a share of a shrinking total.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/warehouse-worker#task-exception-handling" }, { "occupation_slug": "warehouse-worker", "task_id": "receiving-and-quality", "task": "Receiving and checking inbound goods", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "physical; cognitive", "does_not_establish": "Inbound arrives in every state of packaging because suppliers have no incentive to standardise. That incentive changes the moment a large buyer requires it, and large buyers are exactly who automates.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/warehouse-worker#task-receiving-and-quality" }, { "occupation_slug": "warehouse-worker", "task_id": "robot-fleet-operation", "task": "Operating and maintaining the robot fleet", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "physical; cognitive", "does_not_establish": "A fleet needs far fewer technicians than it displaces pickers, and the skills are different enough that the displaced are not the ones hired. This is a real job and a poor answer to the previous three.", "evidence_ids": "ev-20250701-warehouse-worker-1", "url": "https://flyvolo.ai/en/careers/warehouse-worker#task-robot-fleet-operation" }, { "occupation_slug": "assembly-line-worker", "task_id": "repetitive-assembly", "task": "Repetitive assembly at a fixed station", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "Automation decisions here are driven by labour cost relative to capital cost, which is why the same product is assembled by robots in one country and by people in another. Wages, not capability, set the pace in most of the world.", "evidence_ids": "ev-20250925-assembly-line-worker-1; ev-20250925-assembly-line-worker-3", "url": "https://flyvolo.ai/en/careers/assembly-line-worker#task-repetitive-assembly" }, { "occupation_slug": "assembly-line-worker", "task_id": "dexterous-and-variable-assembly", "task": "Dexterous or variable assembly", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical", "does_not_establish": "These stations survived in heavily automated plants, which means they survived as a small remainder inside a plant that already shed most of its line jobs. Surviving the last round is not the same as being safe.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/assembly-line-worker#task-dexterous-and-variable-assembly" }, { "occupation_slug": "assembly-line-worker", "task_id": "inspection-and-quality", "task": "Inspection and quality checks", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Machine vision displaced visual inspection stations in most modern plants. What remains is borderline cases and untrained defect types, which is a fraction of the shifts inspection used to fill.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/assembly-line-worker#task-inspection-and-quality" }, { "occupation_slug": "assembly-line-worker", "task_id": "machine-tending-and-changeover", "task": "Tending machines and running changeovers", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "physical; cognitive", "does_not_establish": "A few people tending several machines is the whole point: the ratio is the saving. This task describes the form the job takes on the way down, not a stable destination.", "evidence_ids": "ev-20251119-assembly-line-worker-2", "url": "https://flyvolo.ai/en/careers/assembly-line-worker#task-machine-tending-and-changeover" }, { "occupation_slug": "assembly-line-worker", "task_id": "robot-programming-and-maintenance", "task": "Programming and maintaining the robots", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "physical; cognitive", "does_not_establish": "Plants report shortages of these technicians while also not funding the training that would produce them from the operators they already have. The opening is real and the bridge to it is not built.", "evidence_ids": "ev-20260702-assembly-line-worker-4", "url": "https://flyvolo.ai/en/careers/assembly-line-worker#task-robot-programming-and-maintenance" }, { "occupation_slug": "electrician", "task_id": "installation-in-buildings", "task": "Installation in real buildings", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical", "does_not_establish": "The least exposed task on the site is still exposed to how many buildings get built. Its risk is the construction cycle and interest rates, not automation — which does not make it a safe decade.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/electrician#task-installation-in-buildings" }, { "occupation_slug": "electrician", "task_id": "fault-finding", "task": "Fault finding", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive; physical", "does_not_establish": "Diagnostic assistants help the less experienced most, which compresses the wage premium that experience used to earn. The task stays human while the reason to pay a veteran for it weakens.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/electrician#task-fault-finding" }, { "occupation_slug": "electrician", "task_id": "compliance-and-certification", "task": "Testing, certification and paperwork", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "A regulatory choice protects the signature, and regulators periodically review what must be signed. Self-certification schemes already exist in several markets.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/electrician#task-compliance-and-certification" }, { "occupation_slug": "electrician", "task_id": "customer-and-scheduling", "task": "Quoting, scheduling and the customer", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Better software here mostly helps the self-employed compete with each other. It raises the floor on admin quality without raising what anyone can charge.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/electrician#task-customer-and-scheduling" }, { "occupation_slug": "electrician", "task_id": "electrification-work", "task": "Electrification: EV chargers, heat pumps, solar, storage", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "physical", "does_not_establish": "This demand is created by policy — subsidies, bans, targets — and policy reverses. A trade whose growth depends on a subsidy schedule should read the schedule.", "evidence_ids": "ev-20250409-electrician-1; ev-20260710-electrician-2", "url": "https://flyvolo.ai/en/careers/electrician#task-electrification-work" }, { "occupation_slug": "chef", "task_id": "fixed-menu-high-volume", "task": "Fixed-menu, high-volume cooking", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "physical; process", "does_not_establish": "Several high-profile robotic kitchen ventures have closed after failing to beat the economics of a low-wage crew. Deployment is real but narrower than the announcements suggested.", "evidence_ids": "ev-20251106-chef-1; ev-20260226-chef-2", "url": "https://flyvolo.ai/en/careers/chef#task-fixed-menu-high-volume" }, { "occupation_slug": "chef", "task_id": "cooking-to-order", "task": "Cooking to order in a real kitchen", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical", "does_not_establish": "Nothing in development runs a Saturday service, but plenty in deployment runs a fast-food line. The protection is specific to restaurants with variable menus, which is a minority of food-service jobs.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/chef#task-cooking-to-order" }, { "occupation_slug": "chef", "task_id": "menu-and-costing", "task": "Menu development and costing", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Costing software raises margins for owners, and most chefs are not owners. The task being assisted does not change who captures the saving.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/chef#task-menu-and-costing" }, { "occupation_slug": "chef", "task_id": "running-the-brigade", "task": "Running the team", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Team management is protected and it is also the reason people leave the industry. A task that is safe because it is punishing is not a reassuring answer.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/chef#task-running-the-brigade" }, { "occupation_slug": "chef", "task_id": "supervising-kitchen-automation", "task": "Working alongside kitchen automation", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "physical; process", "does_not_establish": "Peripheral by weight, and it appears in the kitchens that automated — which are the kitchens with fewer cooks. It is a consequence of the change, not a shelter from it.", "evidence_ids": "ev-20251106-chef-1", "url": "https://flyvolo.ai/en/careers/chef#task-supervising-kitchen-automation" }, { "occupation_slug": "delivery-rider", "task_id": "dispatch-and-routing", "task": "Deciding what to deliver next and which way to go", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "The cognitive half of this job left years ago and nobody counted it as automation at the time. That is worth remembering when judging what is happening to other occupations now.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/delivery-rider#task-dispatch-and-routing" }, { "occupation_slug": "delivery-rider", "task_id": "the-ride", "task": "The ride through the city", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "mobility", "does_not_establish": "Regulation on drones and sidewalk robots is the binding constraint in most cities and could loosen. Watch what your city's authority permits, not what the pilots in other cities announce.", "evidence_ids": "ev-20250422-delivery-rider-1; ev-20251120-delivery-rider-2", "url": "https://flyvolo.ai/en/careers/delivery-rider#task-the-ride" }, { "occupation_slug": "delivery-rider", "task_id": "last-fifty-metres", "task": "The last fifty metres", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "mobility; physical", "does_not_establish": "Being the durable part of the job says nothing about the pay for it. Platforms set per-delivery rates, and a task nobody can automate can still be repriced downward every quarter.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/delivery-rider#task-last-fifty-metres" }, { "occupation_slug": "delivery-rider", "task_id": "handling-the-platform", "task": "Managing the platform and the earnings", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This task is growing because pay structures are getting more complex, which is not a sign of a healthier job. Skill at navigating an opaque system is skill that produces nothing.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/delivery-rider#task-handling-the-platform" }, { "occupation_slug": "delivery-rider", "task_id": "robot-and-drone-support", "task": "Loading, recovering and supervising delivery robots", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "mobility; physical", "does_not_establish": "It exists only where the pilots do, which today is a handful of neighbourhoods. The people-to-robot ratio is a commercial question still being answered, and the answer the operators want is a low one.", "evidence_ids": "ev-20251120-delivery-rider-2", "url": "https://flyvolo.ai/en/careers/delivery-rider#task-robot-and-drone-support" }, { "occupation_slug": "ride-hail-driver", "task_id": "the-drive", "task": "Driving the trip", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "mobility", "does_not_establish": "'Where it is permitted' is most of the story. Deployment covers a small fraction of the world's cities, excludes most weather and road conditions, and has been paused or withdrawn after incidents. Ten years of forecasts have consistently been too fast; treat any timeline, including implied ones, with suspicion.", "evidence_ids": "ev-20231121-ride-hail-driver-4; ev-20260518-ride-hail-driver-3; ev-20260901-ride-hail-driver-1", "url": "https://flyvolo.ai/en/careers/ride-hail-driver#task-the-drive" }, { "occupation_slug": "ride-hail-driver", "task_id": "passenger-handling", "task": "Handling the passenger", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive; physical", "does_not_establish": "Robotaxi operators handle the edges by choosing who may ride and where. Restricting the service is a legitimate product decision, and it removes the need for this task rather than failing at it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/ride-hail-driver#task-passenger-handling" }, { "occupation_slug": "ride-hail-driver", "task_id": "edge-cases-on-the-road", "task": "Roads that are not on the map", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "mobility", "does_not_establish": "Geofences are drawn to exclude these conditions, and geofences expand. The protection is a map boundary, and map boundaries are the thing every operator is working to move.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/ride-hail-driver#task-edge-cases-on-the-road" }, { "occupation_slug": "ride-hail-driver", "task_id": "platform-economics", "task": "Working the platform", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Being good at working the platform is how drivers survive a squeeze, not how they escape one. The site lists it as a task because it consumes real effort, not because it is a skill worth building a career on.", "evidence_ids": "ev-20260816-ride-hail-driver-2", "url": "https://flyvolo.ai/en/careers/ride-hail-driver#task-platform-economics" }, { "occupation_slug": "ride-hail-driver", "task_id": "fleet-support-roles", "task": "Remote assistance and fleet operations", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "mobility; cognitive", "does_not_establish": "Every deployment employs these people today at ratios the operators openly describe as temporary. Planning around a job whose employer has publicly said it intends to need fewer of them is a short plan.", "evidence_ids": "ev-20231121-ride-hail-driver-4", "url": "https://flyvolo.ai/en/careers/ride-hail-driver#task-fleet-support-roles" }, { "occupation_slug": "sales-account-manager", "task_id": "research-and-qualification", "task": "Working out who is worth your time", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Faster qualification is not more deals. It changes where the hours go, not whether the customer buys — and a rep who now qualifies three times as many leads is being measured on a bigger number with the same closing skill behind it.", "evidence_ids": "ev-20250401-sales-account-manager-3; ev-20260812-sales-account-manager-1", "url": "https://flyvolo.ai/en/careers/sales-account-manager#task-research-and-qualification" }, { "occupation_slug": "sales-account-manager", "task_id": "quoting-and-proposals", "task": "Putting the price and the proposal together", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "What automates is assembling the quote, not deciding the number. Where the margin is set by judgement about this customer and this competitor, the document gets faster and the decision does not move — and the decision is where the money is.", "evidence_ids": "ev-20260812-sales-account-manager-1", "url": "https://flyvolo.ai/en/careers/sales-account-manager#task-quoting-and-proposals" }, { "occupation_slug": "sales-account-manager", "task_id": "follow-up-discipline", "task": "Chasing, on time, every time", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "Remembering is not persuading. A perfect follow-up list raises the floor for a disorganised rep and does almost nothing for a good one — so it compresses the gap between reps rather than replacing any of them, and that changes who gets hired, not how many.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/sales-account-manager#task-follow-up-discipline" }, { "occupation_slug": "sales-account-manager", "task_id": "negotiation-and-trust", "task": "The conversation where it is decided", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Being the hardest part to automate does not make it the biggest part of the day. In most SME sales roles this is a few hours a week sitting on top of many hours of assembly and chasing — so a role can shrink substantially while its irreplaceable core stays exactly where it is.", "evidence_ids": "ev-20260528-sales-account-manager-2", "url": "https://flyvolo.ai/en/careers/sales-account-manager#task-negotiation-and-trust" }, { "occupation_slug": "sales-account-manager", "task_id": "pipeline-and-forecast", "task": "Telling the boss what will land", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "A better forecast changes what management does, not what the customer does. It can also make a rep's month look worse without their work being worse, which is a management problem dressed up as a data problem.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/sales-account-manager#task-pipeline-and-forecast" }, { "occupation_slug": "sales-account-manager", "task_id": "operating-the-selling-system", "task": "Running the machine that now sells with you", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "New duties are not new headcount. This lands on the reps who are already there, usually without a title change and without coming off their number — which is how a role gets heavier while looking unchanged on the org chart.", "evidence_ids": "ev-20250401-sales-account-manager-3", "url": "https://flyvolo.ai/en/careers/sales-account-manager#task-operating-the-selling-system" }, { "occupation_slug": "first-line-manager", "task_id": "allocation-and-scheduling", "task": "Deciding who does what today", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "A roster the machine produced still has to be defended to the person who got the bad shift, and that conversation stays with the supervisor. What automates is the hour of arranging, not the accountability for the arrangement — so this role can lose its most visible daily task and keep the one that makes it a job.", "evidence_ids": "ev-20260802-first-line-manager-2", "url": "https://flyvolo.ai/en/careers/first-line-manager#task-allocation-and-scheduling" }, { "occupation_slug": "first-line-manager", "task_id": "spotting-trouble-early", "task": "Seeing it go wrong before the report does", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "More alerts is not more attention. Where detection got cheap, the usual result is a supervisor triaging a longer list of flagged items in the same hours — the work moved from finding to filtering, which reads as no change at all on a job description.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/first-line-manager#task-spotting-trouble-early" }, { "occupation_slug": "first-line-manager", "task_id": "performance-and-feedback", "task": "Writing the review and giving the feedback", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "A drafted review is not a delivered one, and what costs a manager something is saying it out loud to someone who disagrees. It also creates a new failure mode rather than removing one: an assessment assembled from whatever was easy to log quietly promotes whoever is most legible to the system.", "evidence_ids": "ev-20260302-first-line-manager-3; ev-20260514-first-line-manager-1; ev-20260802-first-line-manager-2", "url": "https://flyvolo.ai/en/careers/first-line-manager#task-performance-and-feedback" }, { "occupation_slug": "first-line-manager", "task_id": "keeping-people", "task": "The conversation that keeps someone", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Being the least automatable task does not protect the headcount around it. A company can halve its supervisors with this task fully intact — each remaining one simply does it with twice as many people. That is how an irreplaceable part gets thinner without ever being replaced.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/first-line-manager#task-keeping-people" }, { "occupation_slug": "first-line-manager", "task_id": "reporting-upward", "task": "Telling the floor above what happened", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "What automates is the summary, not the selection. Deciding that a near-miss is worth the boss's attention this week is a judgement with career consequences attached — and a system that reports everything has effectively reported nothing.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/first-line-manager#task-reporting-upward" }, { "occupation_slug": "first-line-manager", "task_id": "answering-for-the-system", "task": "Defending a decision the system made", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "New duties are not new authority. A supervisor asked to explain a system they are not allowed to change carries the blame for it without the power to fix it — a different job from the one they were promoted into, heavier, and invisible on the org chart.", "evidence_ids": "ev-20260302-first-line-manager-3; ev-20260514-first-line-manager-1", "url": "https://flyvolo.ai/en/careers/first-line-manager#task-answering-for-the-system" }, { "occupation_slug": "operations-coordinator", "task_id": "order-to-fulfilment", "task": "Turning a signed order into a real thing happening", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "Orchestration automates the sequence, not the commitment. Someone still has to be willing to promise a customer a date and wear it when the supplier misses — and that promise is worth more than the scheduling. The job can lose most of its clicks and keep all of its liability.", "evidence_ids": "ev-20251231-operations-coordinator-2; ev-20260630-operations-coordinator-1", "url": "https://flyvolo.ai/en/careers/operations-coordinator#task-order-to-fulfilment" }, { "occupation_slug": "operations-coordinator", "task_id": "supplier-coordination", "task": "Getting the confirmation back out of someone upstream", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "It automates the chasing, not the relationship. The reason a supplier squeezes you in on a full day is that they know you, and that favour is not available to an automated reminder. Where supply is tight, the coordinator with the relationship still wins the slot — which protects some coordinators, not the headcount.", "evidence_ids": "ev-20260630-operations-coordinator-1", "url": "https://flyvolo.ai/en/careers/operations-coordinator#task-supplier-coordination" }, { "occupation_slug": "operations-coordinator", "task_id": "exception-handling", "task": "After the plan breaks", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Exceptions are where the value is, and they are also a small share of the hours. A role that is 80% routine coordination and 20% exceptions can lose the 80% and remain a job — for far fewer people, each of them on call more of the time.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/operations-coordinator#task-exception-handling" }, { "occupation_slug": "operations-coordinator", "task_id": "documents-and-compliance", "task": "The paperwork that has to be right", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Automated checking does not move who is liable. When a wrong number gets through, the penalty lands on the company and the explaining lands on the coordinator — so the checking gets cheaper while the exposure stays exactly where it was, and a check that is never wrong is also a check nobody reads any more.", "evidence_ids": "ev-20260630-operations-coordinator-1", "url": "https://flyvolo.ai/en/careers/operations-coordinator#task-documents-and-compliance" }, { "occupation_slug": "operations-coordinator", "task_id": "keeping-the-record-true", "task": "Making the system say what is actually true", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "process", "does_not_establish": "A cleaner record makes the automation above it trustworthy, which is the point — and it also makes the coordinator's contribution invisible, because the work shows up as an absence of problems. That is a bad position to be in when headcount is reviewed.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/operations-coordinator#task-keeping-the-record-true" }, { "occupation_slug": "operations-coordinator", "task_id": "supervising-the-automation", "task": "Watching the automatic flow for quiet mistakes", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Sampling is not coverage, and nobody has decided what rate is enough. This work is usually unfunded and unmeasured, so it is the first thing dropped in a busy week — which is exactly when the automation is running hardest.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/operations-coordinator#task-supervising-the-automation" }, { "occupation_slug": "ai-implementation-lead", "task_id": "choosing-what-to-try-first", "task": "Deciding what to try first", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Hard to automate is not the same as durable. This whole role exists because a transition is under way; when the tools become ordinary the way spreadsheets did, choosing what to automate becomes part of running a function rather than a function of its own. The safety here is the safety of scaffolding.", "evidence_ids": "ev-20250403-ai-implementation-lead-1; ev-20260401-ai-implementation-lead-4", "url": "https://flyvolo.ai/en/careers/ai-implementation-lead#task-choosing-what-to-try-first" }, { "occupation_slug": "ai-implementation-lead", "task_id": "testing-on-our-own-work", "task": "Checking it on this company's own work", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "A good evaluation tells you the tool works on the cases you thought to collect. It says nothing about the cases nobody thought of, and those are where the expensive failures live — so a passing score is a reason to start watching, not a reason to stop.", "evidence_ids": "ev-20250508-ai-implementation-lead-3", "url": "https://flyvolo.ai/en/careers/ai-implementation-lead#task-testing-on-our-own-work" }, { "occupation_slug": "ai-implementation-lead", "task_id": "wiring-into-what-exists", "task": "Wiring it into what is already there", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Cheap integration is why pilots multiply and why most of them are still running as pilots. Being able to connect a tool in an afternoon does not establish that anyone changed how they work, and the two get reported as the same milestone.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/ai-implementation-lead#task-wiring-into-what-exists" }, { "occupation_slug": "ai-implementation-lead", "task_id": "setting-the-guardrails", "task": "Deciding what it may do unsupervised", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "Guardrails being human work does not mean they get written. In most companies nobody has been given this task, so the boundary is settled by default — by whoever configured the tool — and that default is invisible until something goes out that should not have.", "evidence_ids": "ev-20250403-ai-implementation-lead-1; ev-20260802-ai-implementation-lead-2", "url": "https://flyvolo.ai/en/careers/ai-implementation-lead#task-setting-the-guardrails" }, { "occupation_slug": "ai-implementation-lead", "task_id": "getting-people-to-use-it", "task": "Getting people to actually use it", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is the task most likely to be cut, because it is slow and produces no artefact. A company that buys the tools and skips this gets the licence cost and the pilot, and reports both as progress.", "evidence_ids": "ev-20260401-ai-implementation-lead-4", "url": "https://flyvolo.ai/en/careers/ai-implementation-lead#task-getting-people-to-use-it" }, { "occupation_slug": "ai-implementation-lead", "task_id": "answering-when-it-fails", "task": "Being there when it gets something wrong", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Carrying the blame is not the same as holding the authority. This role often answers for decisions it cannot overrule — the budget, the vendor, the deadline were all set elsewhere — and that combination is the most common reason people leave it.", "evidence_ids": "ev-20260802-ai-implementation-lead-2", "url": "https://flyvolo.ai/en/careers/ai-implementation-lead#task-answering-when-it-fails" }, { "occupation_slug": "business-systems-owner", "task_id": "shaping-the-process-in-the-system", "task": "Shaping the process inside the system", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "Faster configuration means more configuration, not less work. Every rule added is a rule that will later be wrong and have to be found — so the role's hours move from building to untangling, and untangling is invisible on a roadmap.", "evidence_ids": "ev-20260401-business-systems-owner-3", "url": "https://flyvolo.ai/en/careers/business-systems-owner#task-shaping-the-process-in-the-system" }, { "occupation_slug": "business-systems-owner", "task_id": "keeping-the-records-meaningful", "task": "Keeping the records worth trusting", "direction": "automating", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Cleaner data makes everything downstream more trustworthy, which is the point — and it also removes the only visible artefact of this job. Nobody notices records that were never wrong, so the work reads as an absence, which is a poor position when the budget is set.", "evidence_ids": "ev-20260802-business-systems-owner-1", "url": "https://flyvolo.ai/en/careers/business-systems-owner#task-keeping-the-records-meaningful" }, { "occupation_slug": "business-systems-owner", "task_id": "deciding-what-the-system-may-decide", "task": "Deciding what the system may decide alone", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Setting the boundary is not the same as being allowed to hold it. This role usually has the knowledge and not the authority: when sales wants the automation to send more, the person who can say no is further up, and being overruled repeatedly is how the boundary erodes without any decision ever being recorded.", "evidence_ids": "ev-20260401-business-systems-owner-3; ev-20260430-business-systems-owner-2; ev-20260802-business-systems-owner-1", "url": "https://flyvolo.ai/en/careers/business-systems-owner#task-deciding-what-the-system-may-decide" }, { "occupation_slug": "business-systems-owner", "task_id": "making-the-numbers-mean-something", "task": "Making the numbers mean something", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Self-service reporting does not remove the person who knows the number is wrong. It multiplies the number of people confidently quoting a figure whose caveat they never saw — so the work shifts from producing reports to correcting them in meetings, which is slower and harder to defend as a headcount.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/business-systems-owner#task-making-the-numbers-mean-something" }, { "occupation_slug": "business-systems-owner", "task_id": "getting-the-team-to-use-it-properly", "task": "Getting the team to use it properly", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This is the first task cut when the role is under pressure, because it produces nothing shippable. Cutting it does not show up as a problem for two quarters, and then shows up as data nobody trusts — by which time the cause is no longer attributable.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/business-systems-owner#task-getting-the-team-to-use-it-properly" }, { "occupation_slug": "partnerships-manager", "task_id": "finding-partners-worth-signing", "task": "Working out which partner is real", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "A longer shortlist is not more partnerships. The bottleneck in this job has never been finding candidates, it is that most signed partners never sell anything — so faster sourcing mostly produces a bigger pile of dormant agreements, and someone still has to explain the pile.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/partnerships-manager#task-finding-partners-worth-signing" }, { "occupation_slug": "partnerships-manager", "task_id": "agreeing-the-terms", "task": "Agreeing what each side actually owes", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Being the hardest part to automate does not make it most of the week. In most partnership roles this is a handful of conversations a year sitting on top of many hours of research, chasing and reporting — so the role can shrink a great deal while the irreplaceable part stays exactly where it is.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/partnerships-manager#task-agreeing-the-terms" }, { "occupation_slug": "partnerships-manager", "task_id": "getting-the-integration-done", "task": "Getting the two systems actually connected", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Cheaper integration is why the number of live connections grows and why most of them carry no volume. A connected partner and a selling partner get reported as the same milestone, and only one of them is revenue.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/partnerships-manager#task-getting-the-integration-done" }, { "occupation_slug": "partnerships-manager", "task_id": "keeping-the-partner-selling", "task": "Keeping them selling after the announcement", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This task surviving does not protect the headcount around it. One manager can nominally cover thirty partners; the work does not disappear, it thins out per partner until the quiet ones are never called — and nobody logs the calls that were not made.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/partnerships-manager#task-keeping-the-partner-selling" }, { "occupation_slug": "partnerships-manager", "task_id": "counting-what-the-channel-produced", "task": "Counting what the channel actually produced", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Automating the count does not settle the argument, it sharpens it. A number that is harder to flatter makes more partnerships look dormant, and the person who has to say so out loud is the same person whose role is justified by the channel being worth having.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/partnerships-manager#task-counting-what-the-channel-produced" }, { "occupation_slug": "partnerships-manager", "task_id": "governing-what-partners-say-in-your-name", "task": "Governing what partners say in your name", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "The obligation moved and the authority did not. Your contract usually lets you object after the fact, not approve in advance — so this is a monitoring duty over someone else's output with no ability to stop it at source, and nobody has decided how much of it to sample.", "evidence_ids": "ev-20260802-partnerships-manager-1", "url": "https://flyvolo.ai/en/careers/partnerships-manager#task-governing-what-partners-say-in-your-name" }, { "occupation_slug": "experienced-software-engineer", "task_id": "changing-a-system-you-did-not-write", "task": "Changing a system you did not write", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "One trial, sixteen developers, mature repositories they knew well — the authors say plainly it does not describe most developers or new projects. It says nothing about how the same tools perform a year later, and a slowdown measured once is not a permanent property of the tools.", "evidence_ids": "ev-20250710-junior-software-developer-2", "url": "https://flyvolo.ai/en/careers/experienced-software-engineer#task-changing-a-system-you-did-not-write" }, { "occupation_slug": "experienced-software-engineer", "task_id": "reviewing-what-the-machine-wrote", "task": "Reviewing what the machine wrote", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A share of code generated says nothing about how long the reviewing takes, whether it is done well, or whether the total number of engineers changed. Reviewing volume rising is also not automatically good work — it is the part of the job most easily done badly under time pressure.", "evidence_ids": "ev-20250415-experienced-software-engineer-3; ev-20250623-experienced-software-engineer-2", "url": "https://flyvolo.ai/en/careers/experienced-software-engineer#task-reviewing-what-the-machine-wrote" }, { "occupation_slug": "experienced-software-engineer", "task_id": "designing-for-how-it-fails", "task": "Designing for how it fails", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is a judgement about the work, not a measurement, and the absence of evidence is the ordinary kind rather than a finding: architecture reviews are not published, so whether design decisions are being handed to a machine or held back from one is not observable from outside the company where it happens.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/experienced-software-engineer#task-designing-for-how-it-fails" }, { "occupation_slug": "experienced-software-engineer", "task_id": "deciding-what-ships", "task": "Deciding what ships", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "A self-reported survey from an interested party is weak evidence for a number and better evidence for a direction. It does not establish that release decisions are becoming harder overall, and it is about large enterprise software rather than all software.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/experienced-software-engineer#task-deciding-what-ships" }, { "occupation_slug": "experienced-software-engineer", "task_id": "making-someone-else-able", "task": "Making someone else able to do it", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Payroll microdata shows an employment decline; it does not show that mentoring increased, or that anyone chose to invest in it. The link between a thinner junior intake and more teaching work falling on seniors is a reading of two facts, not a measured one.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/experienced-software-engineer#task-making-someone-else-able" }, { "occupation_slug": "experienced-software-engineer", "task_id": "owning-the-agent-that-writes", "task": "Owning the agents that write and change code", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "One company's incident and remediation is not a description of the industry, and a remediation programme is evidence that something went wrong rather than evidence about how common it is. Nothing here says this ownership is a role anyone is paid for yet.", "evidence_ids": "ev-20250415-experienced-software-engineer-3", "url": "https://flyvolo.ai/en/careers/experienced-software-engineer#task-owning-the-agent-that-writes" }, { "occupation_slug": "frontend-developer", "task_id": "design-to-interface", "task": "Turning a design into a working interface", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A screen that renders is not a screen that ships. It says nothing about whether the result is maintainable, whether it matches an existing design system, or how it behaves in the states nobody drew — empty, loading, error, half a list.", "evidence_ids": "ev-20260301-frontend-developer-2", "url": "https://flyvolo.ai/en/careers/frontend-developer#task-design-to-interface" }, { "occupation_slug": "frontend-developer", "task_id": "states-nobody-drew", "task": "The states nobody drew", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is a judgement about where the work sits, not a measurement of it. Teams do not count how much of their frontend code is edge-case handling, so there is no figure to check this against, and the balance differs enormously between a marketing page and a trading screen.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/frontend-developer#task-states-nobody-drew" }, { "occupation_slug": "frontend-developer", "task_id": "real-devices-real-networks", "task": "Making it survive real devices and networks", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Nothing here says how many hours this takes or whether teams do it at all — a great deal of shipped frontend work never gets this attention, and the tools existing does not mean they were run.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/frontend-developer#task-real-devices-real-networks" }, { "occupation_slug": "frontend-developer", "task_id": "accessibility-obligation", "task": "Making it work for everyone, because it is required", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "The direction here rests on the shape of the requirement rather than on a measured outcome, and what would settle it is an enforcement action that names a specific interface and what was wrong with it. Requirements differ by market and by whether the product is consumer-facing.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/frontend-developer#task-accessibility-obligation" }, { "occupation_slug": "frontend-developer", "task_id": "owning-a-design-system", "task": "Owning the components everyone else uses", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Whether this is a job or a side duty depends entirely on team size, and how many teams staff it deliberately is not something anyone counts.", "evidence_ids": "ev-20250314-frontend-developer-1", "url": "https://flyvolo.ai/en/careers/frontend-developer#task-owning-a-design-system" }, { "occupation_slug": "frontend-developer", "task_id": "interfaces-that-answer-back", "task": "Building interfaces for things that answer back", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "New work appearing is not the same as new headcount: this is being absorbed into existing roles, and nothing here says anyone was hired to do it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/frontend-developer#task-interfaces-that-answer-back" }, { "occupation_slug": "backend-developer", "task_id": "endpoints-and-plumbing", "task": "Endpoints and the plumbing between them", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Generating an endpoint is not the same as deciding it should exist, what it must guarantee, or what happens when it is called twice. The typing was rarely the expensive part of this work.", "evidence_ids": "ev-20240801-backend-developer-1; ev-20260413-backend-developer-2", "url": "https://flyvolo.ai/en/careers/backend-developer#task-endpoints-and-plumbing" }, { "occupation_slug": "backend-developer", "task_id": "correctness-under-concurrency", "task": "Keeping it correct when things happen at once", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the nature of the work rather than a measured one. Incidents are almost never attributed publicly to generated concurrency code, and that silence is not evidence that it does not happen — a post-incident review naming a cause that specific is unusual in any company.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/backend-developer#task-correctness-under-concurrency" }, { "occupation_slug": "backend-developer", "task_id": "data-model-and-migrations", "task": "The data model, and changing it while it is in use", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Says nothing about how often tools now draft migrations, which they do routinely. The claim is about who decides and answers for it, not about who types it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/backend-developer#task-data-model-and-migrations" }, { "occupation_slug": "backend-developer", "task_id": "the-security-boundary", "task": "Who is allowed to see what", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "This rests on the structure of the problem rather than on a measurement, and what would settle it is a defect study that separates generated code from hand-written code in the same codebase. Regulated environments already require human sign-off here, which may be doing more of the work than the difficulty is.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/backend-developer#task-the-security-boundary" }, { "occupation_slug": "backend-developer", "task_id": "cost-latency-capacity", "task": "What it costs and how fast it is", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Nothing here measures how much of this is now tool-driven in practice, and the answer differs enormously between a team with an observability budget and one without.", "evidence_ids": "ev-20240801-backend-developer-1", "url": "https://flyvolo.ai/en/careers/backend-developer#task-cost-latency-capacity" }, { "occupation_slug": "backend-developer", "task_id": "answering-for-the-service", "task": "Being the one the pager wakes", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "This says nothing about whether on-call load is rising or falling, which is the question most engineers actually care about and the one nobody publishes.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/backend-developer#task-answering-for-the-service" }, { "occupation_slug": "data-engineer", "task_id": "building-the-pipeline", "task": "Building the pipeline", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "The pipeline running is not the same as the numbers being right. Building it was never where the time went in a mature data team; keeping it true was.", "evidence_ids": "ev-20241112-data-engineer-1", "url": "https://flyvolo.ai/en/careers/data-engineer#task-building-the-pipeline" }, { "occupation_slug": "data-engineer", "task_id": "when-the-data-is-wrong", "task": "When the data is wrong and nothing errored", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the nature of the failure rather than a measurement of how often it happens. Silent data errors are not quantified anywhere, partly because by definition they are found late or not at all.", "evidence_ids": "ev-20260128-data-engineer-2", "url": "https://flyvolo.ai/en/careers/data-engineer#task-when-the-data-is-wrong" }, { "occupation_slug": "data-engineer", "task_id": "what-the-number-means", "task": "Owning what a number means", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "That benchmark is about querying rather than about who owns definitions, and its authors sell an alternative system, which the record on the analyst page states. It establishes that enterprise schemas defeat current models; it does not establish that anyone is being hired to maintain semantics.", "evidence_ids": "ev-20241112-data-engineer-1", "url": "https://flyvolo.ai/en/careers/data-engineer#task-what-the-number-means" }, { "occupation_slug": "data-engineer", "task_id": "cost-and-scale", "task": "What it costs to keep asking", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Says nothing about whether costs are rising or falling overall, and cheap querying tends to increase the number of questions asked, which can move the bill in either direction.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/data-engineer#task-cost-and-scale" }, { "occupation_slug": "data-engineer", "task_id": "retention-and-compliance", "task": "What must be kept, deleted, or never collected", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "process", "does_not_establish": "Requirements differ sharply by market and by data type, and what would settle this task is an enforcement action naming a data team's retention or deletion practice rather than the company as a whole — a regulator saying who was supposed to delete what, and by when.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/data-engineer#task-retention-and-compliance" }, { "occupation_slug": "data-engineer", "task_id": "feeding-the-models", "task": "Feeding the systems that answer in sentences", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "New work appearing is not new headcount. In most companies this is being absorbed by whoever already owns the warehouse, and nothing here says otherwise.", "evidence_ids": "ev-20260413-data-engineer-3", "url": "https://flyvolo.ai/en/careers/data-engineer#task-feeding-the-models" }, { "occupation_slug": "product-manager", "task_id": "writing-it-down", "task": "Writing the requirement down", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A well-written spec for the wrong thing is worse than a rough one for the right thing. Producing the document was never the scarce part; knowing which document to write is.", "evidence_ids": "ev-20251125-product-manager-2; ev-20260802-product-manager-1", "url": "https://flyvolo.ai/en/careers/product-manager#task-writing-it-down" }, { "occupation_slug": "product-manager", "task_id": "deciding-what-not-to-build", "task": "Deciding what not to build", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is a judgement about the nature of the work rather than a measured one. Nothing here says organisations value it correctly, and many measure product managers on output rather than on what they prevented.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/product-manager#task-deciding-what-not-to-build" }, { "occupation_slug": "product-manager", "task_id": "finding-the-real-problem", "task": "Finding out what is actually wrong", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Says nothing about how many product managers actually do this. In a great many companies the role never talks to a user, and for those roles this task is theoretical.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/product-manager#task-finding-the-real-problem" }, { "occupation_slug": "product-manager", "task_id": "getting-three-teams-to-move", "task": "Getting three teams to move together", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A description of where the work sits, not a measurement. It also does not say this is done well — coordination failure is among the most common reasons products ship late.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/product-manager#task-getting-three-teams-to-move" }, { "occupation_slug": "product-manager", "task_id": "being-wrong-in-public", "task": "Being wrong in public", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Whether this is real depends entirely on the company. Many product managers have responsibility without the authority that would make it meaningful, and this page cannot tell you which kind a given job is.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/product-manager#task-being-wrong-in-public" }, { "occupation_slug": "product-manager", "task_id": "specifying-what-the-model-may-do", "task": "Specifying what a generative feature may do to a user", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "New work appearing is not new headcount, and nothing here says companies are staffing it. In most teams it is currently absorbed by whoever owns the feature.", "evidence_ids": "ev-20260802-product-manager-1", "url": "https://flyvolo.ai/en/careers/product-manager#task-specifying-what-the-model-may-do" }, { "occupation_slug": "machine-learning-engineer", "task_id": "building-a-model-for-the-problem", "task": "Building a model for the problem", "direction": "automating", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This site's four directions cannot tell those two mechanisms apart, and the difference matters to anyone planning around this page: a task that became purchasable can become unpurchasable again when the vendor's price, licence or capability moves, in a way that a task a machine learned to do does not. It also says nothing about the domains where a bought model is not an option — narrow, proprietary, latency-bound or regulated ones — which are not rare.", "evidence_ids": "ev-20241009-machine-learning-engineer-1", "url": "https://flyvolo.ai/en/careers/machine-learning-engineer#task-building-a-model-for-the-problem" }, { "occupation_slug": "machine-learning-engineer", "task_id": "deciding-good-enough", "task": "Deciding what counts as good enough", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the nature of the work, not a measurement of how it is staffed. Plenty of teams do not do this at all and ship on a vendor's published benchmark, which is the failure this describes rather than evidence against it — but how many is not counted anywhere.", "evidence_ids": "ev-20241009-machine-learning-engineer-1; ev-20260128-machine-learning-engineer-3", "url": "https://flyvolo.ai/en/careers/machine-learning-engineer#task-deciding-good-enough" }, { "occupation_slug": "machine-learning-engineer", "task_id": "data-fit-to-learn-from", "task": "Getting data the thing can learn from", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Nothing here measures how much of a given team's time this takes, and the answer differs by an order of magnitude between a team with an existing data asset and one starting from nothing. Using a model to label data that trains a model also has known failure modes this judgement does not weigh.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/machine-learning-engineer#task-data-fit-to-learn-from" }, { "occupation_slug": "machine-learning-engineer", "task_id": "when-it-quietly-stops-working", "task": "When it quietly stops working", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Says nothing about whether teams actually monitor. A model running unwatched for a year is common and this judgement does not capture it — where nobody watches, this task is not human-led, it is simply not done.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/machine-learning-engineer#task-when-it-quietly-stops-working" }, { "occupation_slug": "machine-learning-engineer", "task_id": "feature-engineering", "task": "Hand-crafting the inputs", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Not evidence about the 2022-2026 window at all; it is older than that and is recorded here for orientation. Domain-specific feature design is also still load-bearing in some regulated settings where an unexplainable input is not allowed, which this judgement does not separate out.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/machine-learning-engineer#task-feature-engineering" }, { "occupation_slug": "machine-learning-engineer", "task_id": "answering-for-what-it-does-to-people", "task": "Answering for what it does to people", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "The record this reasoning points at is not attached to this occupation, and deliberately so: the obligation names the deployer, and attaching it here would claim it lands on engineers when in most organisations nobody has yet been told it lands on them. So this is inference about where the work will sit, not evidence that it already sits there. Requirements also differ sharply by market and by what the system is used for.", "evidence_ids": "ev-20260802-machine-learning-engineer-2", "url": "https://flyvolo.ai/en/careers/machine-learning-engineer#task-answering-for-what-it-does-to-people" }, { "occupation_slug": "radiologist", "task_id": "reading-the-routine-study", "task": "Reading the routine study", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Says nothing about how much of a radiologist's day this is, and the share differs enormously between a screening programme and a tertiary hospital. The two records attached here are a count of what regulators have authorised and a count of what training programmes have funded; neither is deployment. A cleared device is not a hospital running it, and a funded residency post is not a study being read — the first is a ceiling on what is permitted, the second a bet on demand years out. What would settle it: a health system's own report of what share of studies are machine-read before a person sees them.", "evidence_ids": "ev-20161124-radiologist-4; ev-20260526-radiologist-2; ev-20260904-radiologist-1", "url": "https://flyvolo.ai/en/careers/radiologist#task-reading-the-routine-study" }, { "occupation_slug": "radiologist", "task_id": "catching-what-nobody-asked-about", "task": "Catching what nobody asked about", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the structure of the work, not a measurement. How often incidental findings change what happens to a patient is not quantified in a way that transfers, and the answer differs by modality and by population. It also does not claim people are good at this — missed incidentals are a known and studied failure.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/radiologist#task-catching-what-nobody-asked-about" }, { "occupation_slug": "radiologist", "task_id": "deciding-whether-to-image-at-all", "task": "Deciding whether to image at all", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Nothing here says how often radiologists actually protocol studies rather than rubber-stamping them, and in many settings that step has already been delegated or automated away for reasons unrelated to AI.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/radiologist#task-deciding-whether-to-image-at-all" }, { "occupation_slug": "radiologist", "task_id": "the-report-somebody-acts-on", "task": "The report somebody acts on", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Nothing public measures report-drafting time saved, and vendor claims about it are marketing from a party with a stake in the answer. What would settle it: a health system publishing its own before-and-after turnaround times.", "evidence_ids": "ev-20161124-radiologist-4; ev-20250901-radiologist-5", "url": "https://flyvolo.ai/en/careers/radiologist#task-the-report-somebody-acts-on" }, { "occupation_slug": "radiologist", "task_id": "image-guided-procedures", "task": "Procedures done inside a person", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "physical; cognitive", "does_not_establish": "Interventional radiology is really a separate occupation with its own training path, and treating it as one task of this page understates how different it is. It also says nothing about volume, which is where the actual pressure on this work comes from.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/radiologist#task-image-guided-procedures" }, { "occupation_slug": "radiologist", "task_id": "answering-for-the-machine-that-read-it", "task": "Answering for the machine that read it", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "New work appearing is not new headcount, and in most departments this is absorbed by whoever is already there. What would settle it is a hospital publishing a post created for this and the hours attached to it.", "evidence_ids": "ev-20250901-radiologist-5; ev-20260721-radiologist-3; ev-20260904-radiologist-1", "url": "https://flyvolo.ai/en/careers/radiologist#task-answering-for-the-machine-that-read-it" }, { "occupation_slug": "real-estate-agent", "task_id": "knowing-what-is-for-sale", "task": "Knowing what is for sale", "direction": "automating", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Says nothing about markets where listings are not public or not centralised, which is a large share of the world. It also says nothing about whether losing this asset reduced the number of agents — in several markets agent numbers rose while this advantage fell, which is a fact this judgement does not explain.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/real-estate-agent#task-knowing-what-is-for-sale" }, { "occupation_slug": "real-estate-agent", "task_id": "pricing-the-property", "task": "Putting a price on it", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "That record is about one company taking balance-sheet risk on its own valuations in one country during an extraordinary housing market; it does not say automated valuation is inaccurate, and it does not say agents price better. What it establishes is narrower and more useful: an estimate and a commitment are different things, and only the second was scarce.", "evidence_ids": "ev-20211102-real-estate-agent-1", "url": "https://flyvolo.ai/en/careers/real-estate-agent#task-pricing-the-property" }, { "occupation_slug": "real-estate-agent", "task_id": "the-showing", "task": "Standing in the room", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "physical; cognitive", "does_not_establish": "Nothing here measures how many viewings still happen in person, and the answer moved sharply during the pandemic and has partly moved back. It also says nothing about rental, where self-showing has gone much further than in resale.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/real-estate-agent#task-the-showing" }, { "occupation_slug": "real-estate-agent", "task_id": "getting-two-sides-to-agree", "task": "Getting two sides to agree", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the nature of the work, not a measurement of its value. It also does not defend the commission: that this part is human does not establish that a percentage of the sale price is what it is worth, and the fee structure is under pressure in several markets for reasons that have nothing to do with automation.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/real-estate-agent#task-getting-two-sides-to-agree" }, { "occupation_slug": "real-estate-agent", "task_id": "finding-the-next-client", "task": "Finding the next client", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Says nothing about referral, which in most markets is where the majority of business actually comes from and which no tool has moved. It also does not address the oversupply of agents relative to transactions, which is the real pressure on income in several markets and predates any of this.", "evidence_ids": "ev-20251231-real-estate-agent-3", "url": "https://flyvolo.ai/en/careers/real-estate-agent#task-finding-the-next-client" }, { "occupation_slug": "real-estate-agent", "task_id": "answering-for-what-you-said", "task": "Answering for what you told them", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "New work appearing is not new income, and in a commission business unpaid duties are absorbed rather than compensated. Requirements also differ enormously by market, and what would settle the weight of this duty is an enforcement action against an agent over generated material, which has not surfaced in the markets this page covers.", "evidence_ids": "ev-20260401-real-estate-agent-2", "url": "https://flyvolo.ai/en/careers/real-estate-agent#task-answering-for-what-you-said" }, { "occupation_slug": "auditor", "task_id": "sampling-and-testing", "task": "Sampling and testing", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Says nothing about how audit hours are actually billed, and in many firms the hours freed here were absorbed into more testing rather than into fewer people. It also says nothing about smaller audits, where the tooling is often not bought at all.", "evidence_ids": "ev-20250701-auditor-3", "url": "https://flyvolo.ai/en/careers/auditor#task-sampling-and-testing" }, { "occupation_slug": "auditor", "task_id": "deciding-where-to-look", "task": "Deciding where to look", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the nature of the work, not a measurement. It also does not claim auditors are good at this: missing the risk that mattered is the defining failure of this profession and there is a long public record of it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/auditor#task-deciding-where-to-look" }, { "occupation_slug": "auditor", "task_id": "asking-the-question", "task": "Asking the question that gets a real answer", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Nothing here measures how much of a modern audit is inquiry versus document testing, and in many engagements the inquiry is a formality completed by email. Where it is a formality, this task is not human-led — it is simply not being done.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/auditor#task-asking-the-question" }, { "occupation_slug": "auditor", "task_id": "the-file-and-the-opinion", "task": "The file and the opinion", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Nothing public measures documentation time saved in audit, and firm claims about it come from a party selling the transformation. What would settle it: a regulator's inspection findings on files prepared with these tools.", "evidence_ids": "ev-20240701-auditor-2", "url": "https://flyvolo.ai/en/careers/auditor#task-the-file-and-the-opinion" }, { "occupation_slug": "auditor", "task_id": "signing-it", "task": "Signing it", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "It protects the signature, not the headcount behind it. A firm can sign the same number of opinions with fewer people, and the signature says nothing about how many juniors were needed to get there — which is the number most readers of this page actually care about.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/auditor#task-signing-it" }, { "occupation_slug": "auditor", "task_id": "owning-machine-drafted-work", "task": "Owning what the machine drafted", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "That record is one consulting engagement at one firm, not a statutory audit, and it establishes that the failure happened rather than how often it does. New work appearing is also not new headcount: in a billable-hours business an unbilled checking duty is absorbed, which is exactly the condition under which it gets skipped.", "evidence_ids": "ev-20240701-auditor-2; ev-20251007-accountant-2", "url": "https://flyvolo.ai/en/careers/auditor#task-owning-machine-drafted-work" }, { "occupation_slug": "insurance-claims-handler", "task_id": "taking-the-claim", "task": "Taking the claim", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "That figure is one direct-to-consumer insurer writing simple personal lines, and the same filing's next sentence says claims the bot is not authorised to settle, or where it identifies concerns, are routed to human claims experts. So it establishes that the intake end is automated at that company, not that claims work no longer needs people. It also says nothing about traditional insurers running older systems, which is most of the market by premium.", "evidence_ids": "ev-20231231-insurance-claims-handler-1", "url": "https://flyvolo.ai/en/careers/insurance-claims-handler#task-taking-the-claim" }, { "occupation_slug": "insurance-claims-handler", "task_id": "deciding-whether-it-is-covered", "task": "Deciding whether it is covered", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the structure of the work, not a measurement of how often facts are contested — and that share differs enormously by line of business. It also does not say consistent is the same as correct: a rule applied consistently to a badly written policy produces consistent unfairness.", "evidence_ids": "ev-20250806-insurance-claims-handler-2", "url": "https://flyvolo.ai/en/careers/insurance-claims-handler#task-deciding-whether-it-is-covered" }, { "occupation_slug": "insurance-claims-handler", "task_id": "valuing-the-loss", "task": "Putting a number on the loss", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Says nothing about accuracy on unusual items or in markets where parts supply is volatile, and nothing about whether faster estimates changed what claimants receive. Neither has been measured publicly.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/insurance-claims-handler#task-valuing-the-loss" }, { "occupation_slug": "insurance-claims-handler", "task_id": "the-claim-that-does-not-fit", "task": "The claim that does not fit", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about where the risk sits, not a measurement of how many claims are exceptions. It also does not claim humans handle exceptions well — a tired handler at volume is exactly how an expensive claim gets missed.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/insurance-claims-handler#task-the-claim-that-does-not-fit" }, { "occupation_slug": "insurance-claims-handler", "task_id": "telling-someone-no", "task": "Telling someone no", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "It protects the answerability, not the headcount. One person can answer for many automated declines, and in several markets the explanation given is itself templated. What would settle how much the duty weighs is an enforcement action against an insurer over an automated decline, which has not surfaced.", "evidence_ids": "ev-20250806-insurance-claims-handler-2", "url": "https://flyvolo.ai/en/careers/insurance-claims-handler#task-telling-someone-no" }, { "occupation_slug": "insurance-claims-handler", "task_id": "spotting-the-invented-claim", "task": "Spotting the invented claim", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Nothing here measures false positives, which are the whole cost of this task — a wrongly flagged claimant is a person accused of a crime by a statistical model. Fraud-detection performance is also rarely published by insurers, so this rests on the shape of the problem rather than on a measurement.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/insurance-claims-handler#task-spotting-the-invented-claim" }, { "occupation_slug": "counsellor", "task_id": "being-available-at-three-in-the-morning", "task": "Being there at three in the morning", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "It says nothing about whether that support helps, and the trials that would answer that have not reported. It also says nothing about what happens when a low-intensity conversation turns out not to be low-intensity — the case this whole occupation exists for.", "evidence_ids": "ev-20251008-counsellor-3", "url": "https://flyvolo.ai/en/careers/counsellor#task-being-available-at-three-in-the-morning" }, { "occupation_slug": "counsellor", "task_id": "the-notes-and-the-forms", "task": "The notes and the forms", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Notes in this field are clinical records that can be read in a complaint or a court, so a drafted note still has to be read and owned line by line. The time saving claimed for drafting has also not been measured for this profession specifically.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/counsellor#task-the-notes-and-the-forms" }, { "occupation_slug": "counsellor", "task_id": "noticing-what-is-not-said", "task": "Noticing what is not being said", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the structure of the work, not a measurement, and not a claim that practitioners are reliably good at it — missing what mattered is a known and studied failure in this field. It also says nothing about how much of ordinary practice actually reaches this depth.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/counsellor#task-noticing-what-is-not-said" }, { "occupation_slug": "counsellor", "task_id": "the-relationship-itself", "task": "The relationship itself", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is the profession's own account of what works, not something this site can verify, and it is contested within the field. It also does not establish that clients prefer it — for the mildest presentations, some evidently prefer not having a person there at all.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/counsellor#task-the-relationship-itself" }, { "occupation_slug": "counsellor", "task_id": "making-a-therapeutic-decision", "task": "Making a therapeutic decision", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "One US state. It does not establish what other markets will do, and it does not stop products that are marketed as wellness or coaching rather than therapy — which is where most of them already sit. A law also says what may not be offered, not what people actually use at three in the morning.", "evidence_ids": "ev-20250601-counsellor-2; ev-20250801-counsellor-1", "url": "https://flyvolo.ai/en/careers/counsellor#task-making-a-therapeutic-decision" }, { "occupation_slug": "counsellor", "task_id": "answering-for-a-tool-in-the-room", "task": "Answering for a tool in the room", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "New work appearing is not new income, and in private practice an unbilled duty is absorbed. What would settle how hard this bites is a disciplinary action against a practitioner over a tool used in session, which has not surfaced.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/counsellor#task-answering-for-a-tool-in-the-room" }, { "occupation_slug": "government-service-clerk", "task_id": "taking-the-application", "task": "Taking the application", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "process", "does_not_establish": "A policy requiring adoption is not a measurement of adoption, which is why this site weights a mandate cautiously. It says nothing about how many counters actually closed, nothing about the share of applicants who complete online, and nothing about markets without a national platform. Implementation evidence would be a deployment record, not another policy document.", "evidence_ids": "ev-20210511-government-service-clerk-2; ev-20220928-government-service-clerk-1", "url": "https://flyvolo.ai/en/careers/government-service-clerk#task-taking-the-application" }, { "occupation_slug": "government-service-clerk", "task_id": "checking-it-against-the-rule", "task": "Checking it against the rule", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "Says nothing about whether the margin is used well: consistent application of a bad rule produces consistent harm, and discretion at a counter is also where unequal treatment lives. Neither has been measured publicly.", "evidence_ids": "ev-20260721-government-service-clerk-3", "url": "https://flyvolo.ai/en/careers/government-service-clerk#task-checking-it-against-the-rule" }, { "occupation_slug": "government-service-clerk", "task_id": "the-case-the-form-was-not-written-for", "task": "The case the form was not written for", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about where the work sits, not a measurement of how many cases are exceptions. It also does not claim counters handle exceptions well — being told to come back with a document you cannot obtain is the ordinary experience of this failure.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/government-service-clerk#task-the-case-the-form-was-not-written-for" }, { "occupation_slug": "government-service-clerk", "task_id": "the-person-who-cannot-use-the-app", "task": "The person who cannot use the app", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "process", "does_not_establish": "It establishes that the category exists and that policy acknowledges it, not how many counters are actually kept open for it or how well they are staffed. Where a service closes its counter anyway, this task does not become human-led; it stops being done.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/government-service-clerk#task-the-person-who-cannot-use-the-app" }, { "occupation_slug": "government-service-clerk", "task_id": "getting-departments-to-move-together", "task": "Getting departments to move together", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Nothing here says who does this coordination today, and in many places nobody is assigned to: the citizen does it, by visiting four counters. That is the failure this task describes, not evidence against it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/government-service-clerk#task-getting-departments-to-move-together" }, { "occupation_slug": "government-service-clerk", "task_id": "answering-for-what-the-platform-decided", "task": "Answering for what the platform decided", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "New work appearing is not new headcount, and how such challenges are actually handled is not published by any administration this page covers. It also does not establish that a reason is actually available — in several systems the answer a counter can give is the same code the applicant already saw.", "evidence_ids": "ev-20260721-government-service-clerk-3", "url": "https://flyvolo.ai/en/careers/government-service-clerk#task-answering-for-what-the-platform-decided" }, { "occupation_slug": "procurement-specialist", "task_id": "finding-and-comparing", "task": "Finding and comparing suppliers", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Finding a supplier is not qualifying one. Nothing here says the candidates produced are real, solvent, or able to deliver at the volume asked, and in several industries the list that matters is not public at all.", "evidence_ids": "ev-20251005-procurement-specialist-1; ev-20260501-procurement-specialist-3; ev-20260716-procurement-specialist-2", "url": "https://flyvolo.ai/en/careers/procurement-specialist#task-finding-and-comparing" }, { "occupation_slug": "procurement-specialist", "task_id": "the-order-and-the-paperwork", "task": "The order and the paperwork", "direction": "automating", "basis": "inferred", "weight": "core", "technologies": "process", "does_not_establish": "Says nothing about how much of a buyer's week this actually is, and the answer differs enormously between a company with an ERP and one running on spreadsheets — which is most small companies. Automating a match also does not resolve the mismatch, which is where the time goes.", "evidence_ids": "ev-20251005-procurement-specialist-1; ev-20260501-procurement-specialist-3", "url": "https://flyvolo.ai/en/careers/procurement-specialist#task-the-order-and-the-paperwork" }, { "occupation_slug": "procurement-specialist", "task_id": "negotiating-with-someone-who-remembers", "task": "Negotiating with someone who will remember", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about the structure of the work, not a measurement, and it is not an argument that buyers negotiate well. It also does not cover commodity categories bought at auction, where the relationship genuinely does not exist.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/procurement-specialist#task-negotiating-with-someone-who-remembers" }, { "occupation_slug": "procurement-specialist", "task_id": "knowing-a-supplier-is-in-trouble", "task": "Knowing a supplier is in trouble", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Nothing here says whether companies act on what monitoring shows them. A known-risky supplier that nobody replaced is the ordinary outcome, and that is a decision problem rather than a detection one.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/procurement-specialist#task-knowing-a-supplier-is-in-trouble" }, { "occupation_slug": "procurement-specialist", "task_id": "deciding-who-gets-the-short-supply", "task": "Deciding who goes short", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "A judgement about where the decision sits, not a measurement of how often shortages happen — which differs enormously by industry and by year. It also does not claim the decision is made well, or made by procurement rather than above it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/procurement-specialist#task-deciding-who-gets-the-short-supply" }, { "occupation_slug": "procurement-specialist", "task_id": "answering-for-the-supply-chain", "task": "Answering for who is in the chain", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Requirements differ sharply by market and by company size, and what would settle how much the duty weighs is an enforcement action naming a buyer, which has not surfaced. New work appearing is also not new headcount — in most companies this lands on the buyer who already has the supplier relationship.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/procurement-specialist#task-answering-for-the-supply-chain" }, { "occupation_slug": "ai-researcher", "task_id": "writing-the-experiment", "task": "Turning an idea into a running experiment", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Agent-workdays are a measure of effort supplied, not of work replaced — the same disclosure notes that available compute grew substantially over the period, so more experiments does not by itself mean fewer people were needed to run them. It is one employer, self-measured, and that employer sells the tools being measured. Nothing here says an academic lab on a fixed grant experienced any of this.", "evidence_ids": "ev-20260906-ai-researcher-1", "url": "https://flyvolo.ai/en/careers/ai-researcher#task-writing-the-experiment" }, { "occupation_slug": "ai-researcher", "task_id": "keeping-the-rig-running", "task": "Keeping the rig running", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Declining help-desk traffic is consistent with agents answering the questions, and it is equally consistent with the infrastructure having got better, or with the people who used to ask having left. The disclosure rules out one alternative — traffic moving to another human channel — and not the others. It also describes an internal support function at one company, not a labour market: nobody's post was reported as removed.", "evidence_ids": "ev-20260906-ai-researcher-1", "url": "https://flyvolo.ai/en/careers/ai-researcher#task-keeping-the-rig-running" }, { "occupation_slug": "ai-researcher", "task_id": "steering-what-runs-the-experiment", "task": "Steering the thing that runs the experiment", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The intervention rate was produced by an agentic classifier reading session logs — a machine judging machines — which the disclosure states plainly and which no external party has checked. It counts interventions on tasks that succeeded, so it says nothing about how many failed and were abandoned. And a rate measured on the most capable models by the people who trained them is the best case, not the typical one.", "evidence_ids": "ev-20260906-ai-researcher-1; ev-20260906-ai-researcher-3", "url": "https://flyvolo.ai/en/careers/ai-researcher#task-steering-what-runs-the-experiment" }, { "occupation_slug": "ai-researcher", "task_id": "reading-what-a-result-means", "task": "Reading what a result actually means", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is inference about the nature of the work, not a count of anything. It does not establish that teams actually do it well, and the same disclosure notes that the tasks least amenable to automation take on a growing share of researcher effort — which is a claim about where time goes, not evidence that the time is well spent. Nor does it rule out that judging results becomes delegable next; nothing here measures that.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/ai-researcher#task-reading-what-a-result-means" }, { "occupation_slug": "ai-researcher", "task_id": "choosing-what-to-work-on", "task": "Choosing what to work on at all", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A share of tokens is a measure of volume, not of influence: planning is a short activity by nature, so a small token share is what it would look like whether or not machines were doing it. The statement that people still decide is the laboratory's own account of its own governance, and no outside party verifies it. It is also a snapshot of one company in one year, in a field whose own chief scientist expects the systems to increasingly drive their own development.", "evidence_ids": "ev-20260906-ai-researcher-1; ev-20260906-ai-researcher-3", "url": "https://flyvolo.ai/en/careers/ai-researcher#task-choosing-what-to-work-on" }, { "occupation_slug": "ai-researcher", "task_id": "stopping-it", "task": "Stopping it", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "One company's account of one incident, published by that company, with no external audit of what was paused or for how long. A pause is also not a stop: work resumed under stronger controls within weeks, and the same disclosure notes total compute allocation across the analysed workloads was largely unchanged. Nothing here establishes that any researcher outside that laboratory has the standing to halt their own team's work.", "evidence_ids": "ev-20260818-ai-researcher-2; ev-20260912-ai-researcher-4; ev-20260913-ai-researcher-5", "url": "https://flyvolo.ai/en/careers/ai-researcher#task-stopping-it" }, { "occupation_slug": "general-practitioner", "task_id": "writing-the-visit-down", "task": "Writing the visit down", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Time saved on the note is not time given back to the doctor — in most systems it is absorbed by seeing more patients, and whether it lands as relief or as throughput is a management decision made after the tool arrives, not a property of the tool. Nor does it touch what the note is for: a record written to be defensible is a different document from one written to be useful, and automation has so far made the first cheaper without making the second better.", "evidence_ids": "ev-20260801-general-practitioner-1", "url": "https://flyvolo.ai/en/careers/general-practitioner#task-writing-the-visit-down" }, { "occupation_slug": "general-practitioner", "task_id": "working-out-what-is-wrong", "task": "Working out what it is", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A tool being good at the diagnosis does not put it in the room, and in most systems what reaches the consultation is decided by procurement, liability and the electronic record vendor rather than by accuracy. Read the direction as a statement about where the capability is pointing, not as a forecast about your clinic — and note that the same tool arriving can change the job without changing who does it, by turning the doctor into the person who overrides a suggestion and documents why.", "evidence_ids": "ev-20260801-general-practitioner-1", "url": "https://flyvolo.ai/en/careers/general-practitioner#task-working-out-what-is-wrong" }, { "occupation_slug": "general-practitioner", "task_id": "the-examination", "task": "Putting hands on the patient", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical; cognitive", "does_not_establish": "Hard to automate is not the same as valued: in several systems the examination is already the part squeezed hardest by appointment length, and a task can be eroded by the clock without any technology touching it. Remote consultation has removed it entirely from a growing share of visits, and that happened for reasons of cost and access rather than capability.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/general-practitioner#task-the-examination" }, { "occupation_slug": "general-practitioner", "task_id": "deciding-with-the-patient", "task": "Deciding it together, and saying the hard thing", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This being human does not protect the headcount, because it is the part of the visit that scales worst and is therefore the first to be rationed — pushed to a nurse, a leaflet, a follow-up call that does not happen. Read it as a statement about who must do it, not about how many minutes anyone will be given to do it in.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/general-practitioner#task-deciding-with-the-patient" }, { "occupation_slug": "general-practitioner", "task_id": "holding-the-pen", "task": "Holding the prescribing pen", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Protection by regulation is a policy choice, and policy changes: several jurisdictions have already widened who may prescribe, and each widening moved work without removing the pen. The safe thing here is the signature, not the hours behind it — and a system can keep the doctor's name on the prescription while moving the consultation that produced it to somebody cheaper.", "evidence_ids": "ev-20260521-general-practitioner-2", "url": "https://flyvolo.ai/en/careers/general-practitioner#task-holding-the-pen" }, { "occupation_slug": "care-worker", "task_id": "the-body-work", "task": "The body work", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "Difficult to automate has not translated into pay or staffing anywhere we can verify, and treating that difficulty as security gets the economics backwards: this occupation's problem has never been replacement, it has been turnover, injury and a wage set by what a public budget will bear. A task can be irreplaceable and badly paid at the same time, and this one is the clearest example on the site.", "evidence_ids": "ev-20260801-care-worker-1", "url": "https://flyvolo.ai/en/careers/care-worker#task-the-body-work" }, { "occupation_slug": "care-worker", "task_id": "noticing-the-change", "task": "Noticing something changed", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Emerging here does not mean the worker is being replaced on this task; it means the task is being split. The measurement moves to a device and the interpretation stays with a person — but the person doing the interpreting may end up being a remote nurse watching twenty residents rather than the carer in the room, and that is a change in who holds the knowledge, not in whether a machine can care.", "evidence_ids": "ev-20240401-care-worker-2; ev-20260801-care-worker-1", "url": "https://flyvolo.ai/en/careers/care-worker#task-noticing-the-change" }, { "occupation_slug": "care-worker", "task_id": "the-shift-paperwork", "task": "The record and the handover", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "The handover is not the same thing as the chart, and only the chart is being automated. What one carer tells the next about a resident is mostly the part that never enters a form — who is having a bad week, who will refuse a wash from a stranger. Automating the record without noticing that can make the documentation better while making the handover shorter, and the second is where the safety actually lives.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/care-worker#task-the-shift-paperwork" }, { "occupation_slug": "care-worker", "task_id": "being-the-company", "task": "Being the person who is there", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Saying this is human does not mean anyone is paid for it. In most staffing models this task has no line and no minutes assigned to it — it happens in the gaps of the tasks that are counted. A device that fills those gaps does not have to be as good as a person to be chosen, it only has to be cheaper than adding one, and that comparison is made by a budget rather than by a resident.", "evidence_ids": "ev-20240401-care-worker-2", "url": "https://flyvolo.ai/en/careers/care-worker#task-being-the-company" }, { "occupation_slug": "medical-assistant", "task_id": "the-authorisation-chase", "task": "Chasing the authorisation", "direction": "automating", "basis": "inferred", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "Automating one side of an adversarial process does not settle it: payers are automating the denial at the same rate, and the likely outcome is more cycles rather than fewer, with the staff time moving from filling the form to arguing about the reason. It also says nothing about the escalation — the call where somebody explains why this patient needs it now — which is the part that actually gets the yes.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/medical-assistant#task-the-authorisation-chase" }, { "occupation_slug": "medical-assistant", "task_id": "vitals-and-intake", "task": "Vitals and intake", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "process; cognitive", "does_not_establish": "Moving the measurement to a kiosk moves the data but not the responsibility, and the clinic still needs somebody to notice the reading that is wrong rather than merely high. Read this as a task being split rather than removed, and note that where it has been split the observation half has usually not been assigned to anyone.", "evidence_ids": "ev-20260801-medical-assistant-1", "url": "https://flyvolo.ai/en/careers/medical-assistant#task-vitals-and-intake" }, { "occupation_slug": "medical-assistant", "task_id": "the-phone", "task": "The phone", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Human-led here describes who must decide, not how many people get to. Where call volume is the pressure, the usual answer is a script and a longer queue rather than another trained person, and a task can be degraded into a form without being automated at all.", "evidence_ids": "ev-20260801-medical-assistant-1", "url": "https://flyvolo.ai/en/careers/medical-assistant#task-the-phone" }, { "occupation_slug": "medical-assistant", "task_id": "hands-on-procedures", "task": "The hands-on bits", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "physical", "does_not_establish": "This half of the job holding does not make the occupation stable, because the other half is the one that carries the hours. If the administrative work shrinks and the hands-on work does not grow, the role does not disappear — it converts into fewer, more clinical posts, and the entry-level version of this job is the administrative one.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/medical-assistant#task-hands-on-procedures" }, { "occupation_slug": "lab-technician", "task_id": "running-the-analyser", "task": "Running the analyser", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "physical; process", "does_not_establish": "Sixty years of this and the occupation is still here, which makes it the site's clearest counter-example to the idea that automating the core task removes the job. What it removed was the manual method, the entry ladder that taught it and a large share of the headcount per test — the job that remained is supervision of machines, which is a different job at the same title.", "evidence_ids": "ev-20260801-lab-technician-1", "url": "https://flyvolo.ai/en/careers/lab-technician#task-running-the-analyser" }, { "occupation_slug": "lab-technician", "task_id": "knowing-the-result-is-wrong", "task": "Knowing a result is wrong", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "This task being human does not mean it is staffed: laboratories are sized on throughput, and the person who catches the plausible wrong answer produces no measurable output when they succeed. Where lab staffing has been cut, this is the capacity that went, and the consequence appears as a misdiagnosis nobody traces back.", "evidence_ids": "ev-20181201-lab-technician-2; ev-20260801-lab-technician-1", "url": "https://flyvolo.ai/en/careers/lab-technician#task-knowing-the-result-is-wrong" }, { "occupation_slug": "lab-technician", "task_id": "the-specimen", "task": "Handling the specimen", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "physical; process", "does_not_establish": "Deciding a specimen must be recollected is a decision with a cost that lands on somebody else — a patient stuck again, a delayed result — which is why it stays with a person. But the volume of exceptions falls as the pre-analytical process tightens, so this task shrinks for reasons upstream of the laboratory.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/lab-technician#task-the-specimen" }, { "occupation_slug": "lab-technician", "task_id": "the-critical-call", "task": "Making the critical call", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This is protected by protocol rather than by difficulty, and protocols are rewritten. A laboratory under staffing pressure will move to automated notification with acknowledgement tracking, which satisfies the audit and moves the risk to whoever is meant to be reading the queue.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/lab-technician#task-the-critical-call" }, { "occupation_slug": "physical-therapist", "task_id": "the-assessment", "task": "Working out what is actually limiting them", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical; cognitive", "does_not_establish": "Hard to automate and easy to shorten: in session-limited systems the assessment is the part squeezed first, because it produces no treatment the payer recognises. A task can be eroded by a fee schedule without any technology touching it, and in this occupation that has already happened.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/physical-therapist#task-the-assessment" }, { "occupation_slug": "physical-therapist", "task_id": "hands-on-treatment", "task": "Treating with your hands", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "Durable and under pressure from a different direction: several systems are moving away from paying for hands-on treatment toward paying for exercise prescription, on evidence grounds rather than cost grounds. A task can be defended against automation and defunded by a guideline in the same decade.", "evidence_ids": "ev-20250601-physical-therapist-2", "url": "https://flyvolo.ai/en/careers/physical-therapist#task-hands-on-treatment" }, { "occupation_slug": "physical-therapist", "task_id": "the-programme", "task": "Designing the programme", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "A better programme that is not performed is worth nothing, and adherence is the binding constraint in this field rather than programme quality. Automating the design without addressing adherence optimises the half that was not the problem — which is the most common mistake made by products entering this market.", "evidence_ids": "ev-20250601-physical-therapist-2; ev-20260801-physical-therapist-1", "url": "https://flyvolo.ai/en/careers/physical-therapist#task-the-programme" }, { "occupation_slug": "physical-therapist", "task_id": "keeping-them-doing-it", "task": "Keeping them doing it", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is the task with the strongest claim to being the profession's core value and the weakest claim on a payer's schedule: a check-in call is not a billable session in most systems. Where it is unfunded it happens in the therapist's own time, and where it stops happening the outcome falls for reasons no dataset attributes correctly.", "evidence_ids": "ev-20260801-physical-therapist-1", "url": "https://flyvolo.ai/en/careers/physical-therapist#task-keeping-them-doing-it" }, { "occupation_slug": "retail-salesperson", "task_id": "finding-out-what-they-want", "task": "Working out what they actually want", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This holding does not mean the floor keeps its people: the same conversation can be moved to a chat window staffed by fewer people covering more stores, and in that move the task survives while the post does not. What protects this task is proximity to the object, so the exposure rises the moment the store stops holding stock.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/retail-salesperson#task-finding-out-what-they-want" }, { "occupation_slug": "retail-salesperson", "task_id": "knowing-the-stock", "task": "Knowing what is actually in the back", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Losing this task does not free the assistant's time in a way the assistant benefits from, because it was the reason to approach them. It was the errand that started most conversations, and a floor where nobody needs to ask anything is a floor where selling has to start some other way — which nobody has designed.", "evidence_ids": "ev-20260801-retail-salesperson-1", "url": "https://flyvolo.ai/en/careers/retail-salesperson#task-knowing-the-stock" }, { "occupation_slug": "retail-salesperson", "task_id": "the-floor-itself", "task": "Keeping the floor standing up", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "physical; cognitive", "does_not_establish": "Safe from automation and safe from cost are different things. The clearest threat to these hours is not a robot, it is the algorithmic rota: shifts cut to the half hour against predicted footfall, which reduces the same labour without removing a single task. That change has already happened in large chains and shows up in nobody's automation index.", "evidence_ids": "ev-20260801-retail-salesperson-1", "url": "https://flyvolo.ai/en/careers/retail-salesperson#task-the-floor-itself" }, { "occupation_slug": "retail-salesperson", "task_id": "the-transaction", "task": "Taking the money and the return", "direction": "automating", "basis": "inferred", "weight": "peripheral", "technologies": "process", "does_not_establish": "This task moving says little about this occupation, because in most shops it is not where the hours are — cashiering is a separate job here for exactly that reason. Read a self-checkout rollout as a fact about the cashier count, not about whether anyone is left to help you find a size.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/retail-salesperson#task-the-transaction" }, { "occupation_slug": "waiter", "task_id": "taking-the-order", "task": "Taking the order", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "Where this has happened, the covers per server went up and the job got harder rather than smaller — the server now handles more tables and has lost the thirty seconds at the table where they used to read the room. Removing a task can degrade the job it was part of, and this is the cleanest example of it in the service group.", "evidence_ids": "ev-20260701-waiter-1", "url": "https://flyvolo.ai/en/careers/waiter#task-taking-the-order" }, { "occupation_slug": "waiter", "task_id": "running-the-room", "task": "Running the room", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "This being human does not set how many tables one person is given. The same judgement performed across eighteen tables instead of ten is a different job at the same job title, and that ratio is set by a manager after the ordering tablet arrives — which is how automation reaches this task without touching it.", "evidence_ids": "ev-20221221-waiter-2; ev-20260701-waiter-1", "url": "https://flyvolo.ai/en/careers/waiter#task-running-the-room" }, { "occupation_slug": "waiter", "task_id": "carrying", "task": "Carrying things through a crowded room", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "physical; mobility", "does_not_establish": "A running robot in a dining room is as often a marketing decision as a labour one, and where it has been deployed the reported effect is usually on how far staff walk rather than on how many staff there are. Read this direction as the technology working, not as a headcount result — what has been measured is the walking, and no operator has published a service roster before and after.", "evidence_ids": "ev-20221221-waiter-2", "url": "https://flyvolo.ai/en/careers/waiter#task-carrying" }, { "occupation_slug": "waiter", "task_id": "when-it-goes-wrong", "task": "When it goes wrong", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "The authority is the thing, and it is withdrawn more often than it is automated: chains that centralise comp decisions into an app remove this task from the server without a robot being involved at all. What looks like a durable human task can be hollowed out by a policy change in an afternoon.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/waiter#task-when-it-goes-wrong" }, { "occupation_slug": "cleaner", "task_id": "large-floors", "task": "Large open floors", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "physical; mobility", "does_not_establish": "The robot does the easy hectares and leaves the hard square metres, so the work that remains is denser and harder per hour while the hours themselves fall. A crew that loses its open-floor time does not lose its edges, corners, bathrooms or spills — it loses the part of the shift that was recovery.", "evidence_ids": "ev-20260801-cleaner-1", "url": "https://flyvolo.ai/en/careers/cleaner#task-large-floors" }, { "occupation_slug": "cleaner", "task_id": "the-awkward-metres", "task": "The awkward square metres", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical", "does_not_establish": "Being the part that stays is not the same as being the part that is counted. Cleaning contracts are priced on floor area, so the hard metres are usually bundled into a rate set by the easy ones — which means automating the easy part lowers the price of the whole contract without reducing the difficulty of what is left.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/cleaner#task-the-awkward-metres" }, { "occupation_slug": "cleaner", "task_id": "noticing-the-building", "task": "Noticing what the building is doing", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Valuable and unpaid are compatible, and here they have been for a long time. This task is invisible in every contract we are aware of, which means it disappears the moment the contract changes hands — and nobody records that it was lost, because nobody recorded that it existed.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/cleaner#task-noticing-the-building" }, { "occupation_slug": "cleaner", "task_id": "being-scheduled", "task": "Being scheduled and measured", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "This is the change that has actually reached this occupation, and it does not remove a single cleaning task — it removes discretion. Cleaning to a sensor's threshold rather than to a standard changes what the work is, and where it has been introduced the complaint is never that the robot took the job; it is that the route no longer leaves time for anything not on it.", "evidence_ids": "ev-20260801-cleaner-1", "url": "https://flyvolo.ai/en/careers/cleaner#task-being-scheduled" }, { "occupation_slug": "security-guard", "task_id": "watching-screens", "task": "Watching the screens", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Detection is not response, and every deployment we know of converts a watching post into an alert queue rather than removing the post. The important second-order effect is the false-alarm rate: a system that flags too much moves the guard from watching to dismissing, which is a worse job and, after a few hundred dismissals, a less attentive one.", "evidence_ids": "ev-20250601-security-guard-2; ev-20260801-security-guard-1", "url": "https://flyvolo.ai/en/careers/security-guard#task-watching-screens" }, { "occupation_slug": "security-guard", "task_id": "walking-toward-it", "task": "Walking toward the situation", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "physical; cognitive", "does_not_establish": "The task holding does not mean the headcount holds, because the usual design is fewer guards covering more ground with faster alerts. A site can halve its guards and keep response times by adding cameras — the task survives intact and the roster does not, which is the most common shape of change in this occupation.", "evidence_ids": "ev-20250601-security-guard-2; ev-20260801-security-guard-1", "url": "https://flyvolo.ai/en/careers/security-guard#task-walking-toward-it" }, { "occupation_slug": "security-guard", "task_id": "the-door", "task": "Controlling the door", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "process", "does_not_establish": "Automating the rule does not settle who takes responsibility for the exception, and in practice the exception is where every real incident starts. Sites that removed the staffed door and kept only the turnstile have generally re-added a person for reception rather than security reasons — worth knowing, because it means the post can come back under a different name and a lower grade.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/security-guard#task-the-door" }, { "occupation_slug": "security-guard", "task_id": "the-log", "task": "The log and the report", "direction": "automating", "basis": "inferred", "weight": "peripheral", "technologies": "process; cognitive", "does_not_establish": "Automating the log changes what the log is for. Scan points prove presence, not attention, and a route that is verified by scans rewards walking it fast — which is the opposite of what the job needs. That trade-off shows up nowhere in the business case for the system.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/security-guard#task-the-log" }, { "occupation_slug": "it-support-specialist", "task_id": "the-repeat-ticket", "task": "The ticket you have seen four hundred times", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Removing the easy tickets does not leave a smaller version of this job, it leaves a harder one: what remains is the queue's long tail, where the user's description is wrong and the fault is in the gap between two systems. Teams sized on ticket volume that automate the volume and keep the sizing end up with the same headcount doing work the metric no longer describes — and appraisals built on tickets-closed start measuring the wrong thing the week the tool lands.", "evidence_ids": "ev-20190901-it-support-specialist-3; ev-20221001-it-support-specialist-2; ev-20260801-it-support-specialist-1", "url": "https://flyvolo.ai/en/careers/it-support-specialist#task-the-repeat-ticket" }, { "occupation_slug": "it-support-specialist", "task_id": "finding-out-what-actually-happened", "task": "Finding out what actually happened", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Human-led here is about who can solve it, not about how many are employed to. A team that automates the easy half and shrinks by that half leaves the hard half to fewer people, and the hard half is where the burnout in this occupation has always been. The task surviving is not the same as the post surviving.", "evidence_ids": "ev-20260801-it-support-specialist-1", "url": "https://flyvolo.ai/en/careers/it-support-specialist#task-finding-out-what-actually-happened" }, { "occupation_slug": "it-support-specialist", "task_id": "the-desk-visit", "task": "Going to the desk", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "physical", "does_not_establish": "This half being safe from automation is what makes the whole occupation look safer than it is, because the physical half is small and shrinking for reasons unrelated to technology. Measure the ratio in your own week before reading any reassurance into it — in most organisations it is a minority of the hours and falling.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/it-support-specialist#task-the-desk-visit" }, { "occupation_slug": "it-support-specialist", "task_id": "the-quiet-authority", "task": "Being the one who says no", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This is the task most likely to be automated badly rather than well: a verification step that a system performs is a step an attacker can learn exactly. And the authority to refuse is granted, not inherent — a desk measured only on resolution time will grant the access, and no technology is needed for that to happen.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/it-support-specialist#task-the-quiet-authority" }, { "occupation_slug": "cybersecurity-analyst", "task_id": "alert-triage", "task": "Working the alert queue", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The alert queue is where this occupation hires, so automating it removes the training ground rather than the work — the judgement needed further up is currently built by grinding through the queue, and nobody has said what replaces that. Automating triage also does not reduce alerts; it moves the analyst from reading them to tuning what generates them, which is a different and less staffed job.", "evidence_ids": "ev-20221001-cybersecurity-analyst-2; ev-20240124-cybersecurity-analyst-1; ev-20260128-cybersecurity-analyst-3", "url": "https://flyvolo.ai/en/careers/cybersecurity-analyst#task-alert-triage" }, { "occupation_slug": "cybersecurity-analyst", "task_id": "is-this-an-incident", "task": "Deciding it is an incident", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "The decision staying with a person does not mean the person is in your building: this is the task most commonly outsourced to a managed service, which moves it without automating it. Read the direction as being about what kind of thing it is, not as a guarantee that your employer will keep employing someone to do it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/cybersecurity-analyst#task-is-this-an-incident" }, { "occupation_slug": "cybersecurity-analyst", "task_id": "writing-the-detection", "task": "Writing the detection", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Cheap rule-writing makes the false-positive problem worse rather than better, because the constraint was never the writing. A team that can now produce ten times the detections has to be ten times more disciplined about retiring them, and nobody staffs for that — which is how a tool that helps an individual degrades the queue everyone works.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/cybersecurity-analyst#task-writing-the-detection" }, { "occupation_slug": "cybersecurity-analyst", "task_id": "hunting", "task": "Looking for what nothing alerted on", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This is the first activity cut when the queue is loud, because it produces no ticket and no metric. A task can be entirely human and still disappear from a team's week without anybody deciding to remove it — and in this occupation that is the usual way it goes.", "evidence_ids": "ev-20240124-cybersecurity-analyst-1; ev-20250601-cybersecurity-analyst-4", "url": "https://flyvolo.ai/en/careers/cybersecurity-analyst#task-hunting" }, { "occupation_slug": "devops-engineer", "task_id": "writing-the-config", "task": "Writing the configuration", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Faster configuration means more configuration, not less work: every resource created is a resource somebody has to keep, secure and eventually delete, and the hours move from writing to untangling. Untangling is invisible on a roadmap, which is why this is the task most likely to look like a saving and behave like a debt.", "evidence_ids": "ev-20251022-devops-engineer-1; ev-20260122-devops-engineer-3", "url": "https://flyvolo.ai/en/careers/devops-engineer#task-writing-the-config" }, { "occupation_slug": "devops-engineer", "task_id": "the-pager", "task": "Being woken up", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The decision staying human says nothing about how many people are on the rota. The common design is fewer engineers covering more services with better automation, which keeps every task intact and makes the on-call worse — and the sustainability of that is a staffing question that no automation metric captures.", "evidence_ids": "ev-20221001-devops-engineer-2; ev-20251022-devops-engineer-1; ev-20260528-devops-engineer-4", "url": "https://flyvolo.ai/en/careers/devops-engineer#task-the-pager" }, { "occupation_slug": "devops-engineer", "task_id": "what-it-costs", "task": "What it costs and why", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Cheap analysis raises the expectation rather than lowering the work: once a dashboard can name the top ten wasteful resources, somebody has to justify each one that is still there. The task shifts from investigation to explanation, and explanation is a meeting.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/devops-engineer#task-what-it-costs" }, { "occupation_slug": "devops-engineer", "task_id": "designing-the-thing", "task": "Deciding how it should be built", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Being the safest task is also being the smallest: in most teams design decisions are a few days a quarter, and the rest of the week is the work that is being drafted for you. A role can be secure in its most senior task and still lose most of its hours.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/devops-engineer#task-designing-the-thing" }, { "occupation_slug": "technical-writer", "task_id": "drafting-the-reference", "task": "Drafting the reference", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A reference that is generated is a reference nobody read before publishing, and the failure mode is not wrongness but plausibility: documentation that describes what the code declares rather than what it does. That gap is only found by someone using the thing, which is the task below — so automating this raises the value of that one rather than removing the job.", "evidence_ids": "ev-20260401-technical-writer-1; ev-20260721-technical-writer-2", "url": "https://flyvolo.ai/en/careers/technical-writer#task-drafting-the-reference" }, { "occupation_slug": "technical-writer", "task_id": "using-the-thing", "task": "Actually using the thing you are documenting", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This task is the strongest argument for the occupation and the weakest in a budget meeting, because its output is an absence — support tickets that did not happen. Where documentation teams have been cut, this is the task that went first, and nothing about the technology was required for that.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/technical-writer#task-using-the-thing" }, { "occupation_slug": "technical-writer", "task_id": "deciding-what-to-document", "task": "Deciding what does not get written", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Being more valuable is not the same as being recognised: documentation is usually measured by coverage, and coverage is exactly the metric that cheap generation makes meaningless. A team judged on pages published will be rewarded for abandoning this task at the moment it becomes the important one.", "evidence_ids": "ev-20260401-technical-writer-1; ev-20260721-technical-writer-2", "url": "https://flyvolo.ai/en/careers/technical-writer#task-deciding-what-to-document" }, { "occupation_slug": "technical-writer", "task_id": "getting-the-answer", "task": "Getting the answer out of an engineer", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This task is what makes the job hard to do remotely, part-time or from a contractor's chair — which is exactly the direction cost pressure pushes it. The threat here is not automation, it is the job being restructured into something that cannot include this task, after which the documentation degrades for reasons nobody attributes correctly.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/technical-writer#task-getting-the-answer" }, { "occupation_slug": "receptionist", "task_id": "signing-people-in", "task": "Signing people in", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "process", "does_not_establish": "The kiosk handles the compliant visitor and escalates everyone else, and everyone else is the reason the desk exists: the contractor not on the list who is expected, the person who is early and upset, the delivery that cannot be left. Automating the rule does not assign the exception to anyone, and several buildings that removed the desk have re-added a person under a different title and grade.", "evidence_ids": "ev-20250401-receptionist-2; ev-20260801-receptionist-1", "url": "https://flyvolo.ai/en/careers/receptionist#task-signing-people-in" }, { "occupation_slug": "receptionist", "task_id": "the-switchboard", "task": "The phone and the routing", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Routing is not gatekeeping, and the second is what a desk actually does for the people behind it — deciding that this caller should not be put through, or should be put through immediately. Automating the routing without assigning the gatekeeping converts a filter into a funnel, and the cost lands on everyone else's calendar rather than on the desk's budget.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/receptionist#task-the-switchboard" }, { "occupation_slug": "receptionist", "task_id": "whatever-walks-in", "task": "Whatever walks in", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The frequency is the problem. A task that matters enormously and happens twice a week is impossible to justify a full-time post with, which is why this role gets merged into security, facilities or office management rather than removed. The task survives and the job title does not, and that is the most likely shape here.", "evidence_ids": "ev-20250401-receptionist-2", "url": "https://flyvolo.ai/en/careers/receptionist#task-whatever-walks-in" }, { "occupation_slug": "receptionist", "task_id": "the-office-around-it", "task": "The office around the desk", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "This is where the hours actually are for most people with this title, and it is not at the door — which means the exposure of the visible half of the job is a poor guide to the exposure of the post. Anyone reasoning about this occupation from the kiosk alone will get the answer wrong in both directions.", "evidence_ids": "ev-20260801-receptionist-1", "url": "https://flyvolo.ai/en/careers/receptionist#task-the-office-around-it" }, { "occupation_slug": "management-consultant", "task_id": "the-deck", "task": "Building the deck", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The deck was never the product; it was the evidence of effort that justified the fee. What is at risk is therefore the billing model rather than the advice — and the firms most exposed are the ones selling hours, because cheaper production converts directly into less revenue for the same work. Nothing here says the advice gets worse or better.", "evidence_ids": "ev-20250831-management-consultant-2; ev-20260701-management-consultant-1", "url": "https://flyvolo.ai/en/careers/management-consultant#task-the-deck" }, { "occupation_slug": "management-consultant", "task_id": "finding-the-real-question", "task": "Finding the question behind the question", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This being human does not protect the pyramid beneath it. Firms are structured as a small number of people who do this task supported by many who build the deck, and removing the second does not create more of the first — it removes the route by which anyone learned to do it, which is a problem for the firm five years out rather than this quarter.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/management-consultant#task-finding-the-real-question" }, { "occupation_slug": "management-consultant", "task_id": "being-believed", "task": "Being the outside voice", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This is the most durable task in the occupation and the least defensible one to describe out loud, and it does not scale: it is worth a partner's fee for a few conversations, not a team's fees for six months. A firm whose durable value is this cannot be the size it is today.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/management-consultant#task-being-believed" }, { "occupation_slug": "management-consultant", "task_id": "making-it-happen", "task": "Getting three departments to move", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Durable and unpriced at the same time: it is billed as days of a consultant's time, which is exactly the unit that gets cheaper to supply and easier for a client to question when the analysis it used to come bundled with is free. The task survives the technology and may not survive the pricing.", "evidence_ids": "ev-20260701-management-consultant-1", "url": "https://flyvolo.ai/en/careers/management-consultant#task-making-it-happen" }, { "occupation_slug": "loan-officer", "task_id": "scoring-the-application", "task": "Scoring the application", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Sixty years of automating this task did not remove the occupation, and that is the most useful fact on this page — but it did change who gets in and how many. Read a high exposure here as a statement about the ordinary case only: the policy defines what ordinary means, and everything outside it still arrives at a person.", "evidence_ids": "ev-20200204-loan-officer-2; ev-20260801-loan-officer-1", "url": "https://flyvolo.ai/en/careers/loan-officer#task-scoring-the-application" }, { "occupation_slug": "loan-officer", "task_id": "the-exception", "task": "Arguing the exception", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "The exception surviving is not the same as the exception being permitted. Many lenders have narrowed or removed override authority entirely, which deletes this task without automating it — and where that has happened the stated reason is consistency rather than cost. A task can be irreplaceable and still be abolished by policy.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/loan-officer#task-the-exception" }, { "occupation_slug": "loan-officer", "task_id": "gathering-the-case", "task": "Gathering the file", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "This is the task that employs the junior half of a lending team, so automating it removes the entry route rather than the work — the judgement needed above it was learned by handling a few hundred files. Nobody has said what replaces that, and the consequence arrives as a shortage of experienced officers five years later.", "evidence_ids": "ev-20260801-loan-officer-1", "url": "https://flyvolo.ai/en/careers/loan-officer#task-gathering-the-case" }, { "occupation_slug": "loan-officer", "task_id": "the-signature", "task": "Being the name on the file", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "A signature is cheaper to keep than a decision, and a bank can keep the name on the file while moving everything that produced it. Human review that consists of clicking approve on a model's output is human review in name only, and no measurement distinguishes the two from outside.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/loan-officer#task-the-signature" }, { "occupation_slug": "compliance-officer", "task_id": "reading-the-rule", "task": "Reading the rule", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Faster reading raises the expected scope rather than reducing the work: once summarising is free, the question becomes why you have not assessed every jurisdiction. The hours move from reading to defending an interpretation, which is a meeting rather than a document.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/compliance-officer#task-reading-the-rule" }, { "occupation_slug": "compliance-officer", "task_id": "the-alert-queue", "task": "Working the alerts", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "The queue is where this profession hires and where the pattern recognition is learned, so automating it removes the training ground. And the regulator's expectation is that alerts are investigated, not that they are closed — a system that closes more of them faster produces a metric that improves while the risk does not.", "evidence_ids": "ev-20241201-compliance-officer-1", "url": "https://flyvolo.ai/en/careers/compliance-officer#task-the-alert-queue" }, { "occupation_slug": "compliance-officer", "task_id": "saying-no", "task": "Saying no to the business", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Held in place by structure, not by difficulty, and structure is cheap to change: a compliance function that reports into the business it polices has this task on paper and not in practice. The question to ask is not whether a tool could do it but who you report to, and that answer changes without any technology involved.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/compliance-officer#task-saying-no" }, { "occupation_slug": "compliance-officer", "task_id": "the-evidence-file", "task": "Proving it was done", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "Automating the evidence does not automate the judgement it evidences, and it creates a specific new risk: a complete record of a control that was not actually effective is worse than an incomplete one, because it is harder to challenge. Auditors have begun asking how the record was produced, which is a question the tooling was not designed to answer.", "evidence_ids": "ev-20241201-compliance-officer-1; ev-20260721-compliance-officer-2", "url": "https://flyvolo.ai/en/careers/compliance-officer#task-the-evidence-file" }, { "occupation_slug": "construction-worker", "task_id": "building-in-place", "task": "Building it where it goes", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "The real mechanism moving this work is not a robot arriving on site, it is the work leaving the site: prefabrication moves the same task into a factory where a fixed reference frame exists, and the job that results is a different job in a different place with different pay. Reading this task as safe because robots are bad on site misses where the change actually comes from.", "evidence_ids": "ev-20240401-construction-worker-2; ev-20260801-construction-worker-1", "url": "https://flyvolo.ai/en/careers/construction-worker#task-building-in-place" }, { "occupation_slug": "construction-worker", "task_id": "setting-out", "task": "Setting out and checking", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "Better measurement finds more errors and does not fix any of them, so the work it creates is rework, which is manual. Where scanning has been introduced the reported effect is usually fewer buried mistakes rather than fewer people, and the cost saved lands with the client rather than with the crew.", "evidence_ids": "ev-20240401-construction-worker-2", "url": "https://flyvolo.ai/en/careers/construction-worker#task-setting-out" }, { "occupation_slug": "construction-worker", "task_id": "the-sequence", "task": "Working out what can happen today", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Being unautomatable is not being paid: this judgement is usually performed by a working foreman whose pay reflects the trade rather than the coordination. And the same task is being centralised in large contractors, which moves it to an office without automating it at all.", "evidence_ids": "ev-20260801-construction-worker-1", "url": "https://flyvolo.ai/en/careers/construction-worker#task-the-sequence" }, { "occupation_slug": "construction-worker", "task_id": "not-getting-hurt", "task": "Not getting hurt", "direction": "emerging", "basis": "inferred", "weight": "core", "technologies": "cognitive; physical", "does_not_establish": "Safety monitoring changes who is blamed as much as who is safe: a system that records every violation creates a record that can be used to attribute an accident to a worker rather than to a schedule. Where this has been introduced without changing the schedule, the reported effect on injury rates has been much smaller than the reported effect on documentation.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/construction-worker#task-not-getting-hurt" }, { "occupation_slug": "auto-mechanic", "task_id": "reading-the-fault", "task": "Reading the fault code", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "A code names a symptom, not a cause, and the gap between the two is where the trade lives — the same code can mean a failed sensor, a chafed wire or a mouse. Better codes have made the easy jobs faster without touching the hard ones, and the commercial consequence has been to compress the billable time on the easy jobs rather than to reduce the number of mechanics.", "evidence_ids": "ev-20200401-auto-mechanic-2; ev-20260701-auto-mechanic-1", "url": "https://flyvolo.ai/en/careers/auto-mechanic#task-reading-the-fault" }, { "occupation_slug": "auto-mechanic", "task_id": "the-intermittent-fault", "task": "The fault that will not repeat", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive; physical", "does_not_establish": "Being the skilled half does not mean being the paid half: workshops bill by book time, and book time is set for the standard job. The intermittent fault is where the trade's expertise lives and also where its unbilled hours live, which is a pricing problem no technology created and none will solve.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/auto-mechanic#task-the-intermittent-fault" }, { "occupation_slug": "auto-mechanic", "task_id": "doing-the-repair", "task": "Doing the repair", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "The task is safe and the volume is not. Electric drivetrains have far fewer serviceable parts and far fewer scheduled services, so the threat to this trade is the number of jobs arriving rather than who performs them — a change driven by what is sold, not by what can be automated.", "evidence_ids": "ev-20200401-auto-mechanic-2; ev-20260701-auto-mechanic-1", "url": "https://flyvolo.ai/en/careers/auto-mechanic#task-doing-the-repair" }, { "occupation_slug": "auto-mechanic", "task_id": "telling-the-customer", "task": "Telling the customer what it needs", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Trust is the trade's asset and also what makes it vulnerable to a change in ownership rather than in technology: workshops consolidating into chains replace the person who will be there next year with a process, and the customer notices this long before any tool is involved.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/auto-mechanic#task-telling-the-customer" }, { "occupation_slug": "ecommerce-operator", "task_id": "the-listing", "task": "Writing the listing", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Cheap listings do not create advantage, they remove one: when everybody's copy improves at once the ranking returns to what it was and the work still has to be done. This is the clearest case on the site of automation that is compulsory rather than optional — a merchant who does not adopt it falls behind, and one who does gains nothing durable.", "evidence_ids": "ev-20250314-ecommerce-operator-2; ev-20251016-ecommerce-operator-1", "url": "https://flyvolo.ai/en/careers/ecommerce-operator#task-the-listing" }, { "occupation_slug": "ecommerce-operator", "task_id": "buying-traffic", "task": "Buying the traffic", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "This automation was not adopted, it was imposed, and that distinction matters more here than anywhere else on the site: the operator's judgement did not lose to a better algorithm, it lost to a change in what the platform exposes. What is left is choosing what to sell and when to stop — decisions the platform's tools do not make because they are not in the platform's interest to make.", "evidence_ids": "ev-20251016-ecommerce-operator-1", "url": "https://flyvolo.ai/en/careers/ecommerce-operator#task-buying-traffic" }, { "occupation_slug": "ecommerce-operator", "task_id": "reading-the-numbers", "task": "Reading what the numbers mean", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This task holding does not mean the headcount does, because the same person can now cover several stores: cheaper listings and automated bidding raise the number of storefronts one operator can hold, which reduces the operators per store without removing a single task from any of them.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/ecommerce-operator#task-reading-the-numbers" }, { "occupation_slug": "ecommerce-operator", "task_id": "the-rules-changed", "task": "The platform changed the rules again", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This is the most durable task and the least transferable: it is knowledge of one platform, and it becomes worthless the day the merchant moves to another. An occupation whose defensible skill is specific to a single company's rulebook has a different kind of fragility from the automation this page otherwise describes.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/ecommerce-operator#task-the-rules-changed" }, { "occupation_slug": "train-driver", "task_id": "driving-between-stations", "task": "Driving between stations", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "mobility; process", "does_not_establish": "What moved is the control, not the cab. On the grade most of the world's metro kilometres actually run at, the machine drives and a person still sits in front — so the hours did not go anywhere, the content of them changed. It also says nothing about how fast this spreads, because what gates it is civil engineering: sealed track and platform doors, paid for over a decade, not software bought in a quarter.", "evidence_ids": "ev-20231231-train-driver-1; ev-20240301-train-driver-3", "url": "https://flyvolo.ai/en/careers/train-driver#task-driving-between-stations" }, { "occupation_slug": "train-driver", "task_id": "the-doors", "task": "Closing the doors", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "process", "does_not_establish": "A wall is not a machine learning to look, so nothing here transfers to any other occupation. It also does not establish that the person watching the platform disappeared: on lines with screen doors the watching moved to cameras in a control room, and the control room is staffed.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/train-driver#task-the-doors" }, { "occupation_slug": "train-driver", "task_id": "when-the-train-stops-in-a-tunnel", "task": "When the train stops in a tunnel", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "process", "does_not_establish": "Being irreplaceable in an emergency is not the same as being employed for one. These events are rare by design, and an operator can keep the capability without keeping the roster — by putting fewer people on more trains, or by keeping licensed drivers on call rather than in cabs. The protection is real and it is thin.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/train-driver#task-when-the-train-stops-in-a-tunnel" }, { "occupation_slug": "train-driver", "task_id": "being-someone-on-the-system", "task": "Being a person who is there", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "Staff presence in the system is not the same job as driving a train, it usually pays differently, and it is rostered differently. Counting it as the driver surviving would be the mistake this page exists to avoid — what survived is a headcount in a company, not a seat in a cab, and the person filling it may be someone else.", "evidence_ids": "ev-20260522-train-driver-2", "url": "https://flyvolo.ai/en/careers/train-driver#task-being-someone-on-the-system" }, { "occupation_slug": "train-driver", "task_id": "the-control-room", "task": "Running the line from a screen", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "process; cognitive", "does_not_establish": "New work appearing is not new headcount, and here the arithmetic is unusually visible: one desk watches a line that used to need a driver on every train. It is also not the same career. The route from a cab to a control desk exists but it is not automatic, and the same workforce study that named this function also noted that the people in the cabs would themselves be measured by sensors on the trains.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/train-driver#task-the-control-room" }, { "occupation_slug": "train-driver", "task_id": "taking-it-back-by-hand", "task": "Taking the train back by hand", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "mobility", "does_not_establish": "A reserve is not a roster. Keeping the capability requires far fewer licensed people than running the service used to, and it requires them to stay current on a skill they now rarely use, which is a training cost somebody has to carry. This also describes one country's chosen path; nothing here says other operators will take it rather than going straight to unattended operation.", "evidence_ids": "ev-20240301-train-driver-3", "url": "https://flyvolo.ai/en/careers/train-driver#task-taking-it-back-by-hand" }, { "occupation_slug": "port-worker", "task_id": "horizontal-transport", "task": "Moving the box between quay and yard", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "mobility; physical", "does_not_establish": "A task removed at one terminal is not a task removed at a port, and the same company still recruits for the seat at its other terminals. It also says nothing about the total: when demand surged, the same operator reactivated conventional berths and ramped up manpower, so the conventional terminals and the people who crew them are the capacity that gets switched on. The work exists, and it exists as a buffer.", "evidence_ids": "ev-20251231-port-worker-1", "url": "https://flyvolo.ai/en/careers/port-worker#task-horizontal-transport" }, { "occupation_slug": "port-worker", "task_id": "crane-work", "task": "Working the crane", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "physical; process", "does_not_establish": "Eighty per cent automated is not eighty per cent of the job gone: the remaining fifth is the part that needs a person, and one person can now hold several cranes. It also does not say where the person sits. In one market the cab is being kept by a contract clause rather than by the technology, which means the direction here is partly a bargaining outcome, and bargaining outcomes expire.", "evidence_ids": "ev-20241001-port-worker-2", "url": "https://flyvolo.ai/en/careers/port-worker#task-crane-work" }, { "occupation_slug": "port-worker", "task_id": "lashing", "task": "Lashing the boxes down", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical", "does_not_establish": "Being hard to automate is not the same as being well paid or safe, and a job that survives because it is unpleasant is surviving on a thin reason. The operator ran tabletop exercises for abnormal lashing scenarios with nearly a hundred participants in the same year it automated the handoff, which tells you the risk is live rather than resolved.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/port-worker#task-lashing" }, { "occupation_slug": "port-worker", "task_id": "keeping-the-fleet-running", "task": "Keeping the fleet running", "direction": "emerging", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "New work appearing is not the same headcount, the same pay or the same person. These are engineering and information technology roles with their own entry requirements, and nothing in the operator's own account says that the people whose seat was removed are the people who filled them. It also concentrates: one control room covers a terminal that used to need a driver in every vehicle.", "evidence_ids": "ev-20251231-port-worker-1", "url": "https://flyvolo.ai/en/careers/port-worker#task-keeping-the-fleet-running" }, { "occupation_slug": "port-worker", "task_id": "what-the-contract-says", "task": "What the contract says a machine may do", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "A contract covers the ports its parties operate and expires on a date. It is also not evidence that the technology failed: it is evidence that where dockworkers had the leverage to bargain over it, they wrote the manning into the agreement. In ports without that leverage the same equipment runs differently, which is why this page cannot be generalised from any one market.", "evidence_ids": "ev-20241001-port-worker-2", "url": "https://flyvolo.ai/en/careers/port-worker#task-what-the-contract-says" }, { "occupation_slug": "content-moderator", "task_id": "the-queue", "task": "Working the queue", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "A share of actions is not a share of jobs, and this platform does not publish moderator headcount alongside it. It also says nothing about the hard half: an automated system that clears 94% of a queue may be clearing the easy 94%, leaving a residue that is slower and heavier per item than the old average. Nothing here measures that, and anyone claiming the remaining work is proportionally smaller is going beyond the document.", "evidence_ids": "ev-20260630-content-moderator-1", "url": "https://flyvolo.ai/en/careers/content-moderator#task-the-queue" }, { "occupation_slug": "content-moderator", "task_id": "the-borderline-call", "task": "The call the rule does not decide", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "No platform publishes how many items reach this route, and that share is exactly what would decide whether this is a shrinking specialism or a growing one. A judgement that a task is human-led is a judgement about what the work requires, not a prediction that the headcount holds.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/content-moderator#task-the-borderline-call" }, { "occupation_slug": "content-moderator", "task_id": "the-appeal", "task": "Reviewing an appeal", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Compulsory is not the same as staffed, and a right to appeal says nothing about how long the queue is or who is in it. This page has no measurement of appeal volumes or of how many are upheld.", "evidence_ids": "ev-20250630-content-moderator-2", "url": "https://flyvolo.ai/en/careers/content-moderator#task-the-appeal" }, { "occupation_slug": "content-moderator", "task_id": "writing-the-rule", "task": "Turning a value into an enforceable line", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Nothing on this page measures how many people do this at any platform, and it is plausibly a small number even where the queue is large. A task being hard to automate does not make it a common job.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/content-moderator#task-writing-the-rule" }, { "occupation_slug": "content-moderator", "task_id": "teaching-the-system", "task": "Correcting the system that replaced the queue", "direction": "emerging", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "New work is not necessarily equivalent work: this page holds no evidence about how many such roles exist, what they pay, or whether they are offered to the people whose queue was automated. Naming a task that appeared is not a claim that it absorbs the people displaced.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/content-moderator#task-teaching-the-system" }, { "occupation_slug": "content-moderator", "task_id": "the-worst-of-it", "task": "Looking at the worst of it", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "That direction is an inference from the shape of the split, not a measurement: no platform publishes what proportion of human-reviewed items are severe, and none is obliged to.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/content-moderator#task-the-worst-of-it" }, { "occupation_slug": "air-traffic-controller", "task_id": "separation", "task": "Keeping aircraft apart", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "This judgement is about the job as the regulator currently designs it, not about what the technology could do. The thing that would change it fastest is not a capability announcement but a change in who is allowed to hold the certificate — and that change would arrive as a rule, which is checkable in advance, unlike a capability claim.", "evidence_ids": "ev-20150220-air-traffic-controller-2", "url": "https://flyvolo.ai/en/careers/air-traffic-controller#task-separation" }, { "occupation_slug": "air-traffic-controller", "task_id": "the-sequence", "task": "Building the sequence", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "The document states an intention, not an installation. It names no facility, no date and no measured effect on how a sequence is built today, so it does not establish that any of it is in use. What it does establish is which task the agency itself considers the one worth pointing a model at — and that is a different and weaker claim than deployment.", "evidence_ids": "ev-20250326-air-traffic-controller-3", "url": "https://flyvolo.ai/en/careers/air-traffic-controller#task-the-sequence" }, { "occupation_slug": "air-traffic-controller", "task_id": "the-unexpected", "task": "When the plan breaks", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Rare and critical is not the same as safe: a task can stay with people and still support far fewer of them, if the rest of the day is thinned out around it. The workforce plan does not report how often this task is reached, and without that figure the share of a shift it accounts for is unknown.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/air-traffic-controller#task-the-unexpected" }, { "occupation_slug": "air-traffic-controller", "task_id": "training-the-next-one", "task": "Training the next one", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "A simulator shortens the path to certification; it does not establish that the final sign-off moves off a person, and the plan does not claim that. The number of facilities is a deployment commitment rather than a completed count, and nothing here measures how much live instruction time it actually displaced.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/air-traffic-controller#task-training-the-next-one" }, { "occupation_slug": "air-traffic-controller", "task_id": "staffing-the-watch", "task": "Getting the watch covered", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "One employer's stated practice is one employer's stated practice. It says nothing about other providers, and the same document names optimising scheduling efficiency as one of three strategic pillars — so this is a description of the present that the author is arguing against, which makes it unusually credible and unusually likely to change. What it cannot tell you is when.", "evidence_ids": "ev-20250930-air-traffic-controller-1", "url": "https://flyvolo.ai/en/careers/air-traffic-controller#task-staffing-the-watch" }, { "occupation_slug": "air-traffic-controller", "task_id": "new-entrants", "task": "Making room for traffic that has no pilot", "direction": "emerging", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive; physical", "does_not_establish": "Preparing for something is not doing it, and the plan gives no measurement of how much of anyone's shift this currently occupies — at most facilities the honest answer is none. A task labelled emerging is a claim that the work is being created, not a claim that it is already a job.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/air-traffic-controller#task-new-entrants" }, { "occupation_slug": "airline-pilot", "task_id": "hand-flying", "task": "Flying it by hand", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive; physical", "does_not_establish": "A currency rule is a floor, not a description of the day: three landings in ninety days is compatible with a great deal of the rest being flown by the autopilot, and on a long-haul roster it often is. The rule also accepts a full flight simulator in place of the aircraft, so even this requirement is partly met by a machine. What it establishes is that the authorisation is tied to a person who has recently done it, not how much of the flight that person flew.", "evidence_ids": "ev-20111103-airline-pilot-1", "url": "https://flyvolo.ai/en/careers/airline-pilot#task-hand-flying" }, { "occupation_slug": "airline-pilot", "task_id": "managing-the-automation", "task": "Working the automation", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "Supervising is not obviously less work than flying, and nothing on this page measures whether it is. There is also a known cost that this direction does not capture: skill fade on the manual task underneath, which is part of why the currency rule exists at all. A task labelled augmenting says the machine does more of it than before; it does not say the person's day got easier.", "evidence_ids": "ev-20260610-airline-pilot-2", "url": "https://flyvolo.ai/en/careers/airline-pilot#task-managing-the-automation" }, { "occupation_slug": "airline-pilot", "task_id": "the-abnormal", "task": "When two systems disagree", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "Rare and critical does not mean numerous. A task can stay with people while the number of people needed to cover it falls, if the rest of the flight thins out around it — which is precisely what reduced-crew proposals are about. Nothing here counts how often this task is reached, and the regulation does not report it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/airline-pilot#task-the-abnormal" }, { "occupation_slug": "airline-pilot", "task_id": "the-commanders-decision", "task": "The decision that is signed", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "An accountability rule is a rule, and rules change: this judgement would flip if liability were reassigned, and that would arrive as a legislative proposal rather than as a product. It also says nothing about how much judgement is left in the decision itself once every input arrives pre-computed, which is a different question and one this page does not answer.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/airline-pilot#task-the-commanders-decision" }, { "occupation_slug": "airline-pilot", "task_id": "staying-current", "task": "Keeping the licence alive", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "A simulator substituting for an aircraft in a currency check is not the same as a simulator substituting for flying, and the regulation does not claim it is. Nor does any of this say how much recurrent training a pilot actually does, or whether the total is rising or falling — the rule sets a minimum and operators set the rest.", "evidence_ids": "ev-20111103-airline-pilot-1", "url": "https://flyvolo.ai/en/careers/airline-pilot#task-staying-current" }, { "occupation_slug": "airline-pilot", "task_id": "the-new-category", "task": "A type rating for an aircraft that did not exist", "direction": "emerging", "basis": "evidenced", "weight": "peripheral", "technologies": "cognitive; mobility", "does_not_establish": "A category existing in a regulation is not the same as people being employed in it, and this page holds nothing that counts how many are. Rule-making runs ahead of fleets on purpose, so the honest reading is that the door has been defined, not that anyone is through it.", "evidence_ids": "ev-20241121-airline-pilot-3", "url": "https://flyvolo.ai/en/careers/airline-pilot#task-the-new-category" }, { "occupation_slug": "bus-driver", "task_id": "the-drive", "task": "Driving the route", "direction": "automating", "basis": "evidenced", "weight": "core", "technologies": "mobility", "does_not_establish": "The approved driverless routes are short and slow — one in Japan runs about 630 metres at up to about 12 km/h — and every case on this page is a pilot rather than a network. Nothing here establishes that a full-size bus can do a busy urban route without someone watching it, and the authorities involved say plainly that they have not concluded that either.", "evidence_ids": "ev-20240531-bus-driver-3; ev-20251002-bus-driver-1; ev-20260304-bus-driver-2", "url": "https://flyvolo.ai/en/careers/bus-driver#task-the-drive" }, { "occupation_slug": "bus-driver", "task_id": "getting-people-on", "task": "Getting people on and off safely", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "physical; cognitive", "does_not_establish": "Keeping a person on board is a pilot plan, not a permanent rule, and nothing here says whether it survives the move from pilot to network. A role that is kept is not necessarily kept for everyone who holds it now.", "evidence_ids": "ev-20231210-bus-driver-4; ev-20251002-bus-driver-1", "url": "https://flyvolo.ai/en/careers/bus-driver#task-getting-people-on" }, { "occupation_slug": "bus-driver", "task_id": "the-passenger-in-trouble", "task": "The passenger in trouble", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive; physical", "does_not_establish": "That a person is needed does not say which person, how many, or at what pay. A remote operator can call for help but cannot kneel beside someone, and this page holds no evidence about how often incidents like these happen per route.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/bus-driver#task-the-passenger-in-trouble" }, { "occupation_slug": "bus-driver", "task_id": "safety-operator", "task": "Watching the bus drive itself", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "mobility; cognitive", "does_not_establish": "A role designed to end is not a career path by itself. Watching an automated system for hours and intervening rarely is known to be hard for people to do well, and nothing here measures how the pilots handle that.", "evidence_ids": "ev-20231210-bus-driver-4; ev-20251002-bus-driver-1", "url": "https://flyvolo.ai/en/careers/bus-driver#task-safety-operator" }, { "occupation_slug": "bus-driver", "task_id": "remote-supervision", "task": "Supervising several buses from a control room", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "The ratio is the whole economic question and it is not settled: a published illustration of one operator to three vehicles is a design figure, not a measured one. Fewer control-room seats than driver seats is the point of the arrangement, so this task does not absorb everyone whose driving it replaces.", "evidence_ids": "ev-20240531-bus-driver-3; ev-20251002-bus-driver-1", "url": "https://flyvolo.ai/en/careers/bus-driver#task-remote-supervision" }, { "occupation_slug": "bus-driver", "task_id": "when-the-plan-breaks", "task": "When the route does not go to plan", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "mobility; cognitive", "does_not_establish": "Who does this in a driverless network — a roving technician, a nearby driver, a recovery contractor — is not decided by anything on this page, and it may not be a bus driver at all.", "evidence_ids": "ev-20240531-bus-driver-3", "url": "https://flyvolo.ai/en/careers/bus-driver#task-when-the-plan-breaks" }, { "occupation_slug": "actuary", "task_id": "the-data", "task": "Preparing the data and the experience study", "direction": "automating", "basis": "inferred", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "This rests on how the work is done, not on a measurement of how much of it is automated at any insurer. Faster preparation does not make the assumptions behind it any less a person's responsibility, and a data problem found late is still found by someone.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/actuary#task-the-data" }, { "occupation_slug": "actuary", "task_id": "the-price", "task": "Building the pricing model", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "These are expectations about how pricing must be governed, not measurements of how pricing work is now divided between people and software. They also apply to regulated insurers in two jurisdictions; elsewhere the obligations differ.", "evidence_ids": "ev-20240711-actuary-1; ev-20250806-actuary-2", "url": "https://flyvolo.ai/en/careers/actuary#task-the-price" }, { "occupation_slug": "actuary", "task_id": "the-reserves", "task": "Setting the reserves", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "Naming who is responsible for the controls does not say how much of the calculation is done by software, and the supervisor's opinion is addressed to national authorities rather than directly to firms.", "evidence_ids": "ev-20250806-actuary-2", "url": "https://flyvolo.ai/en/careers/actuary#task-the-reserves" }, { "occupation_slug": "actuary", "task_id": "fairness-testing", "task": "Testing a model for unfair discrimination", "direction": "emerging", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "An expectation that testing happens is not evidence of who does it; at some insurers it sits with data science or compliance rather than actuaries. And the tension between group fairness and actuarial fairness is named by the regulator, not resolved by it.", "evidence_ids": "ev-20240711-actuary-1; ev-20250806-actuary-2", "url": "https://flyvolo.ai/en/careers/actuary#task-fairness-testing" }, { "occupation_slug": "actuary", "task_id": "the-opinion", "task": "Signing off on the numbers", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A duty that stays with a named function says nothing about how many people that function employs, and a smaller team signing off on more automated work is entirely consistent with it.", "evidence_ids": "ev-20250806-actuary-2", "url": "https://flyvolo.ai/en/careers/actuary#task-the-opinion" }, { "occupation_slug": "actuary", "task_id": "explaining-it", "task": "Explaining the number to the people who decide", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Nothing here says actuaries in particular are the ones asked to explain, rather than data scientists or risk managers; who explains a model to a regulator varies by insurer and has not been measured.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/actuary#task-explaining-it" }, { "occupation_slug": "civil-engineer", "task_id": "the-calculations", "task": "Running the design calculations", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This rests on how the work is done, not on a measurement of how much calculation software does at any firm. A faster calculation is not a checked one, and nothing here says the checking gets any smaller.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/civil-engineer#task-the-calculations" }, { "occupation_slug": "civil-engineer", "task_id": "the-model", "task": "Producing the drawings and the building model", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive; process", "does_not_establish": "A digital submission mandate is not evidence that the model builds itself, and in Singapore it arrives in phases by project size. It also says nothing about who inside a firm builds the model — engineers or specialist modellers.", "evidence_ids": "ev-20250910-civil-engineer-1; ev-20260723-civil-engineer-3", "url": "https://flyvolo.ai/en/careers/civil-engineer#task-the-model" }, { "occupation_slug": "civil-engineer", "task_id": "the-submission", "task": "Getting the plans approved", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "process", "does_not_establish": "The regulator reports time savings of up to two months on the projects that have used it, but that is its own account rather than an independent measurement, and the system is still being phased in, now only for larger projects. It changes the route the submission takes, not who is responsible for what is submitted.", "evidence_ids": "ev-20250910-civil-engineer-1; ev-20260723-civil-engineer-3", "url": "https://flyvolo.ai/en/careers/civil-engineer#task-the-submission" }, { "occupation_slug": "civil-engineer", "task_id": "signing-the-plans", "task": "Signing the plans as the qualified person", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "A duty to sign does not say how many engineers a firm needs; one qualified person can sign for work produced by fewer people with better tools. This page cites Singapore's law, and other countries assign the duty in their own ways.", "evidence_ids": "ev-20260701-civil-engineer-2", "url": "https://flyvolo.ai/en/careers/civil-engineer#task-signing-the-plans" }, { "occupation_slug": "civil-engineer", "task_id": "the-independent-check", "task": "Checking someone else's structural design", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "That the law relies on a person's certificate says nothing about how the checker does the work, which may well involve independent software models. It is also a senior role held by relatively few engineers. The same section also keeps the regulator's right to carry out random checks on the plans and calculations before approving.", "evidence_ids": "ev-20260701-civil-engineer-2", "url": "https://flyvolo.ai/en/careers/civil-engineer#task-the-independent-check" }, { "occupation_slug": "civil-engineer", "task_id": "on-site", "task": "Deciding on site when reality differs from the drawing", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive; physical", "does_not_establish": "This rests on the nature of site work rather than on a record of how site decisions are made, and it would change if regulators began accepting remote or sensor-based sign-off of site changes.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/civil-engineer#task-on-site" }, { "occupation_slug": "firefighter", "task_id": "the-attack", "task": "Attacking the fire at close range", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "A robot that goes in first does not show that fewer people go in at all, and the service's own figure on that is a forecast, not a measurement. This also says nothing about services that have not bought such robots, which is most of them.", "evidence_ids": "ev-20211115-firefighter-2; ev-20220601-firefighter-1", "url": "https://flyvolo.ai/en/careers/firefighter#task-the-attack" }, { "occupation_slug": "firefighter", "task_id": "the-long-fire", "task": "Holding a large fire over hours", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "physical", "does_not_establish": "The report lists fires where the machine was used; it does not say how many firefighters it replaced at those fires, and at the two where it gives a number, about 50 and about 70 firefighters were deployed. It is one service's record, not a measure of practice elsewhere.", "evidence_ids": "ev-20220601-firefighter-1", "url": "https://flyvolo.ai/en/careers/firefighter#task-the-long-fire" }, { "occupation_slug": "firefighter", "task_id": "search-and-rescue", "task": "Searching the building and getting people out", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical", "does_not_establish": "This rests on what the robots in service are built to do, not on a record of rescues. A casualty-carrying robot is in development at the same service, and if it enters routine use this judgement should be revisited.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/firefighter#task-search-and-rescue" }, { "occupation_slug": "firefighter", "task_id": "command-on-scene", "task": "Reading the fire and commanding the scene", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "The same service describes a long-range goal of robots coordinating among themselves under one operator; that is a stated vision, not a deployment, and nothing here shows it happening.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/firefighter#task-command-on-scene" }, { "occupation_slug": "firefighter", "task_id": "the-robot-operator", "task": "Deploying and operating the robots", "direction": "emerging", "basis": "inferred", "weight": "significant", "technologies": "physical; cognitive", "does_not_establish": "The same source expects robots to be deployed autonomously from vehicles in future, which would shrink the handling part of this task; how large the operating role stays is not known.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/firefighter#task-the-robot-operator" }, { "occupation_slug": "firefighter", "task_id": "prevention", "task": "Prevention, drills and community work", "direction": "human_led", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive", "does_not_establish": "This rests on the nature of the work rather than on a record, and training itself is changing — the same service is building simulators of mixed-use buildings, road tunnels and underground stations to train in.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/firefighter#task-prevention" }, { "occupation_slug": "cabin-crew", "task_id": "the-evacuation", "task": "Being ready to evacuate the aircraft", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "A legal minimum is not a forecast of staffing: it says the floor cannot fall below a count of seats and exits, not how many crew an airline will choose to carry above it. Regulators can change these rules, so the floor is only as fixed as the rule.", "evidence_ids": "ev-19650309-cabin-crew-2; ev-20181001-cabin-crew-1", "url": "https://flyvolo.ai/en/careers/cabin-crew#task-the-evacuation" }, { "occupation_slug": "cabin-crew", "task_id": "crew-in-charge", "task": "Leading the cabin as crew-in-charge", "direction": "human_led", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "The rule defines who is accountable, not how the work of leading a crew is done; scheduling, briefing material and reporting may all be increasingly digital.", "evidence_ids": "ev-20181001-cabin-crew-1", "url": "https://flyvolo.ai/en/careers/cabin-crew#task-crew-in-charge" }, { "occupation_slug": "cabin-crew", "task_id": "the-service", "task": "Serving food and drink", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "physical", "does_not_establish": "This rests on what current cabins and trolleys are, not on a measurement of how widely ordering has been digitised. If service is automated, the legal minimum stays, but crew numbers above it could fall.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/cabin-crew#task-the-service" }, { "occupation_slug": "cabin-crew", "task_id": "passenger-care", "task": "Handling medical and security incidents", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical; cognitive", "does_not_establish": "This rests on the nature of the work rather than on a record, and remote support tools are changing how decisions about diversion and treatment are made.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/cabin-crew#task-passenger-care" }, { "occupation_slug": "cabin-crew", "task_id": "pre-flight-checks", "task": "Checking the cabin and the emergency equipment", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "physical; process", "does_not_establish": "How much of this is digital varies by airline and aircraft type, and nothing here measures it.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/cabin-crew#task-pre-flight-checks" }, { "occupation_slug": "cabin-crew", "task_id": "sales-and-paperwork", "task": "Onboard sales and flight paperwork", "direction": "automating", "basis": "inferred", "weight": "peripheral", "technologies": "process", "does_not_establish": "Removing these steps saves time within a flight; it does not change the legal minimum crew, which is set for safety, not for sales.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/cabin-crew#task-sales-and-paperwork" }, { "occupation_slug": "police-officer", "task_id": "the-patrol", "task": "Patrolling and watching public places", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "mobility; physical", "does_not_establish": "A drone that patrols does not show that fewer officers patrol; the force says the systems optimise manpower deployment, not that they reduce headcount. The unmanned vessel at sea is described as a trial, replacing one manned patrol twice a week, which is not yet a deployment.", "evidence_ids": "ev-20260522-police-officer-1", "url": "https://flyvolo.ai/en/careers/police-officer#task-the-patrol" }, { "occupation_slug": "police-officer", "task_id": "taking-reports", "task": "Taking reports from the public", "direction": "augmenting", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This covers reports lodged at kiosks; reports made to an officer in person or by phone are not described, and nothing here says how many reports now go through the chatbot.", "evidence_ids": "ev-20260522-police-officer-1; ev-20260522-police-officer-2", "url": "https://flyvolo.ai/en/careers/police-officer#task-taking-reports" }, { "occupation_slug": "police-officer", "task_id": "the-investigation", "task": "Investigating a case", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "The minister described the case summary module as a pilot implementation, with further modules due next year and one still in development, so this is a pilot, not yet a deployment.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/police-officer#task-the-investigation" }, { "occupation_slug": "police-officer", "task_id": "traffic-cases", "task": "Processing traffic violations from video", "direction": "automating", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "The force's own account the same day describes the system as one it is exploring, while the minister called it progressively rolling out. Because the two differ on how far along it is, this page does not treat it as deployed; neither says how many cases it handles.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/police-officer#task-traffic-cases" }, { "occupation_slug": "police-officer", "task_id": "the-response", "task": "Responding to incidents and making arrests", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "physical; cognitive", "does_not_establish": "This rests on the legal powers of the role rather than on a record. The same speech mentions armed drones in development for special operations, which is a different and narrower use and is not evidence about ordinary response.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/police-officer#task-the-response" }, { "occupation_slug": "police-officer", "task_id": "community", "task": "Working with the community", "direction": "human_led", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive", "does_not_establish": "This rests on the nature of the work rather than on a record of how outreach is done.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/police-officer#task-community" }, { "occupation_slug": "farmer", "task_id": "the-spraying", "task": "Spraying crops", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "physical; mobility", "does_not_establish": "The areas are the ministry's own estimates, made by methods that changed between years (and not counted at all for one year), so they show scale, not a reliable growth rate. Cumulative sprayed area counts the same field more than once, and nothing here says how many farmers use drones or hire a spraying service instead.", "evidence_ids": "ev-20250331-farmer-1", "url": "https://flyvolo.ai/en/careers/farmer#task-the-spraying" }, { "occupation_slug": "farmer", "task_id": "field-work", "task": "Ploughing, planting and harvesting with machines", "direction": "augmenting", "basis": "inferred", "weight": "core", "technologies": "mobility; physical", "does_not_establish": "A law that funds adoption does not show adoption happened; how many farmers have been certified under it, and what they bought, is not established here.", "evidence_ids": "ev-20241001-farmer-2", "url": "https://flyvolo.ai/en/careers/farmer#task-field-work" }, { "occupation_slug": "farmer", "task_id": "the-decisions", "task": "Deciding what to plant, when and how", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This rests on the nature of the work rather than on a record, and decision-support services are spreading, particularly through the service providers the same ministry promotes.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/farmer#task-the-decisions" }, { "occupation_slug": "farmer", "task_id": "hand-harvest", "task": "Picking and handling fruit and vegetables by hand", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "physical", "does_not_establish": "This rests on the state of the technology rather than on a record, and the same law also certifies and supports companies that develop and supply smart-farming technology, so it may change for specific crops.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/farmer#task-hand-harvest" }, { "occupation_slug": "farmer", "task_id": "the-livestock", "task": "Caring for livestock", "direction": "augmenting", "basis": "evidenced", "weight": "peripheral", "technologies": "physical", "does_not_establish": "A share of milk produced by robots shows the machines are in use, not how much of a farmer's day they take over; the daily health checks and the work when something goes wrong are not measured by it.", "evidence_ids": "ev-20211231-farmer-3", "url": "https://flyvolo.ai/en/careers/farmer#task-the-livestock" }, { "occupation_slug": "farmer", "task_id": "farm-business", "task": "Running the farm's paperwork and sales", "direction": "augmenting", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive; process", "does_not_establish": "This rests on the tools available rather than on a record of how widely they are used.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/farmer#task-farm-business" }, { "occupation_slug": "social-worker", "task_id": "case-notes", "task": "Writing up sessions as case notes", "direction": "automating", "basis": "evidenced", "weight": "significant", "technologies": "cognitive", "does_not_establish": "Agencies using a tool is not the same as every session being recorded with it, and the minister's figure counts agencies, not social workers or sessions. It also says nothing about how accurate the notes are or how much checking they need.", "evidence_ids": "ev-20260326-social-worker-1", "url": "https://flyvolo.ai/en/careers/social-worker#task-case-notes" }, { "occupation_slug": "social-worker", "task_id": "spotting-need", "task": "Spotting families who need help early", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This is funding for pilots and early-stage experiments, not a deployed system, so it does not show that any family is being identified this way today.", "evidence_ids": "ev-20260702-social-worker-2", "url": "https://flyvolo.ai/en/careers/social-worker#task-spotting-need" }, { "occupation_slug": "social-worker", "task_id": "the-relationship", "task": "Working with clients over time", "direction": "human_led", "basis": "inferred", "weight": "core", "technologies": "cognitive", "does_not_establish": "This rests on the nature of the work and on the ministry's stated intent, not on a measurement of whether social workers now spend more time with clients.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/social-worker#task-the-relationship" }, { "occupation_slug": "social-worker", "task_id": "risk-and-crisis", "task": "Assessing risk and responding to crises", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive; physical", "does_not_establish": "These are stated principles for adopting technology, not a rule that forbids a tool from scoring risk, and nothing here shows how risk assessments are actually made today.", "evidence_ids": "ev-20240401-social-worker-3", "url": "https://flyvolo.ai/en/careers/social-worker#task-risk-and-crisis" }, { "occupation_slug": "social-worker", "task_id": "coordinating-help", "task": "Connecting clients to services", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "process", "does_not_establish": "This rests on the systems in place rather than on a record of how referrals now work or how long they take.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/social-worker#task-coordinating-help" }, { "occupation_slug": "social-worker", "task_id": "reports", "task": "Writing reports and applications", "direction": "augmenting", "basis": "inferred", "weight": "peripheral", "technologies": "cognitive; process", "does_not_establish": "This rests on what current tools can do rather than on a record of report-writing tools in use.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/social-worker#task-reports" }, { "occupation_slug": "dentist", "task_id": "reading-xrays", "task": "Reading dental X-rays", "direction": "augmenting", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "The performance figures come from the manufacturer's own submission, on 352 images, and describe readers in a study rather than dentists in practice. A clearance shows the tool may be sold for this use, not how many practices use it.", "evidence_ids": "ev-19991015-dentist-1; ev-20220510-dentist-2", "url": "https://flyvolo.ai/en/careers/dentist#task-reading-xrays" }, { "occupation_slug": "dentist", "task_id": "diagnosis", "task": "Diagnosing and planning treatment", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "cognitive", "does_not_establish": "These are the conditions of one tool's clearance and one country's licensing law; they fix who is responsible, not how much a dentist relies on the tool in practice.", "evidence_ids": "ev-19991015-dentist-1", "url": "https://flyvolo.ai/en/careers/dentist#task-diagnosis" }, { "occupation_slug": "dentist", "task_id": "the-procedures", "task": "Carrying out treatment", "direction": "human_led", "basis": "evidenced", "weight": "core", "technologies": "physical", "does_not_establish": "This rests on the law and the state of the technology rather than on a record of how treatment is delivered, and robotic assistance exists for some specialised procedures such as implant placement.", "evidence_ids": "ev-19991015-dentist-1", "url": "https://flyvolo.ai/en/careers/dentist#task-the-procedures" }, { "occupation_slug": "dentist", "task_id": "records", "task": "Charting and clinical records", "direction": "augmenting", "basis": "inferred", "weight": "significant", "technologies": "cognitive; process", "does_not_establish": "This rests on the capability of the tools rather than on a record of how widely they are used for charting.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/dentist#task-records" }, { "occupation_slug": "dentist", "task_id": "the-patient", "task": "Explaining and reassuring", "direction": "human_led", "basis": "inferred", "weight": "significant", "technologies": "cognitive", "does_not_establish": "This rests on the nature of the work rather than on a record, and it would change if patients came to accept treatment plans explained by software without a dentist present.", "evidence_ids": "", "url": "https://flyvolo.ai/en/careers/dentist#task-the-patient" }, { "occupation_slug": "dentist", "task_id": "supervision", "task": "Supervising the dental team", "direction": "human_led", "basis": "evidenced", "weight": "peripheral", "technologies": "cognitive", "does_not_establish": "The supervision condition applies to some categories of therapist and has exceptions, including public service; other countries organise the dental team differently.", "evidence_ids": "ev-19991015-dentist-1", "url": "https://flyvolo.ai/en/careers/dentist#task-supervision" } ], "evidence": [ { "id": "ev-19650309-cabin-crew-2", "occupation_slug": "cabin-crew", "task_ids": "the-evacuation", "title": "US federal rules require one flight attendant per 50 seats above 100, seated near the exits for take-off and landing", "stage": "constraint", "occurred_on": "1965-03-09", "verified_on": "2026-09-23", "source_name": "Electronic Code of Federal Regulations — 14 CFR 121.391, Flight attendants", "source_url": "https://www.ecfr.gov/current/title-14/chapter-I/subchapter-G/part-121/subpart-M/section-121.391", "source_tier": "primary", "scope": "The current text of 14 CFR 121.391 on the government's own eCFR, originally published at 30 FR 3206 on 9 March 1965 and amended since. It requires one flight attendant on aeroplanes with more than 9 or 19 (by payload) but fewer than 51 seats, two for 51 to 100 seats, and two plus one for each unit or part of a unit of 50 seats above 100. If an operator's emergency evacuation demonstration used more attendants than that, it may not later fly the aeroplane with fewer. During take-off and landing, the required flight attendants must sit as near as practicable to the required floor-level exits and be uniformly distributed through the aeroplane, to provide the most effective egress of passengers in an emergency evacuation. It establishes a legal minimum and its stated safety purpose in the United States; it does not say how many attendants airlines roster above it, and it says nothing about onboard service.", "url": "https://flyvolo.ai/en/changes/ev-19650309-cabin-crew-2" }, { "id": "ev-19991015-dentist-1", "occupation_slug": "dentist", "task_ids": "diagnosis; the-procedures; supervision; reading-xrays", "title": "Singapore's Dental Registration Act reserves the practice of dentistry, including radiographic work, to registered dentists with a practising certificate", "stage": "constraint", "occurred_on": "1999-10-15", "verified_on": "2026-09-24", "source_name": "Singapore Statutes Online (Attorney-General's Chambers) — Dental Registration Act 1999, sections 2, 21 and 29, current version", "source_url": "https://sso.agc.gov.sg/Act/DRA1999", "source_tier": "primary", "scope": "The Act as published on Singapore Statutes Online, in force since 15 October 1999, read in the current version (2022 revised edition). Section 2 defines the practice of dentistry to include any procedure on and treatment of the teeth, jaws or associated structures; radiographic work in connection with them; administering an anaesthetic for such procedures; and procedures for dentures and dental appliances. Section 29(1): a person must not practise dentistry in Singapore unless the person is a registered dentist and has in force a practising certificate; registered oral health therapists may practise within their prescribed scope, and contravention is an offence. Section 21(4) makes it a condition of registration for several categories of oral health therapist that they practise only under the supervision of a registered dentist, with exceptions including the public service. It establishes who may lawfully do this work in Singapore; it says nothing about the tools a dentist uses or how many dentists there are.", "url": "https://flyvolo.ai/en/changes/ev-19991015-dentist-1" }, { "id": "ev-20111103-airline-pilot-1", "occupation_slug": "airline-pilot", "task_ids": "hand-flying; staying-current", "title": "EU law bars a pilot from operating a commercial flight unless they personally carried out three take-offs, approaches and landings in the preceding 90 days", "stage": "constraint", "occurred_on": "2011-11-03", "verified_on": "2026-09-22", "source_name": "EUR-Lex, consolidated text of Commission Regulation (EU) No 1178/2011 (Part-FCL)", "source_url": "https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:02011R1178-20260430", "source_tier": "primary", "scope": "Commission Regulation (EU) No 1178/2011 laying down technical requirements and administrative procedures related to civil aviation aircrew, read in its consolidated version in force from 30 April 2026. The provision is FCL.060, Recent experience: a pilot shall not operate an aircraft in commercial air transport or for carrying passengers as pilot-in-command or co-pilot unless he or she has carried out, in the preceding 90 days, at least three take-offs, approaches and landings as pilot flying in an aircraft of the same type or class, or in a full flight simulator representing that type or class. Two things follow and they point in opposite directions, which is why this record links two tasks rather than one. The requirement is written around a named person having personally done the thing, which is a rarer shape than a rule about who may operate a machine. And the same sentence accepts a simulator in place of the aircraft, which is a machine standing in for the very activity the rule exists to preserve. What this establishes is the shape of the obligation, not a measurement: it says nothing about how much of a given flight is flown by hand, how many pilots are employed, or whether crew composition will change - that last would arrive as a separate rulemaking, proposed in public with a comment period. The regulation speaks only for the European regime; other authorities impose requirements of similar shape and different numbers.", "url": "https://flyvolo.ai/en/changes/ev-20111103-airline-pilot-1" }, { "id": "ev-20150220-air-traffic-controller-2", "occupation_slug": "air-traffic-controller", "task_ids": "separation", "title": "EU law requires air traffic control for a specific sector to be provided by a named licence holder endorsed for that sector, with competence re-assessed before the endorsement expires", "stage": "constraint", "occurred_on": "2015-02-20", "verified_on": "2026-09-22", "source_name": "EUR-Lex, consolidated text of Commission Regulation (EU) 2015/340", "source_url": "https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:02015R0340-20260513", "source_tier": "primary", "scope": "Commission Regulation (EU) 2015/340 on air traffic controller licences, read in its consolidated version in force from 13 May 2026. Two provisions carry the judgement. ATCO.B.005: holders of an air traffic controller licence are authorised to provide air traffic control services in accordance with the ratings and rating endorsements of their licence. ATCO.B.020(a): the unit endorsement authorises the licence holder to provide air traffic control services for a specific sector, group of sectors or working positions under the responsibility of an air traffic services unit, and renewal requires that competence has been assessed under the unit competence scheme within the three months before expiry. What this establishes is the shape of the gate rather than a measurement of anything: in European airspace the authorisation to work a sector attaches to a person, is specific to that sector, and lapses without re-assessment. What it does not establish is anything about how much of the surrounding work a machine does, how many controllers are employed, or whether this rule will hold - it is a regulation, so a change would be proposed in public before taking effect. The regulation is also not a statement about the world outside the European Union, and providers elsewhere are licensed under their own rules.", "url": "https://flyvolo.ai/en/changes/ev-20150220-air-traffic-controller-2" }, { "id": "ev-20161124-radiologist-4", "occupation_slug": "radiologist", "task_ids": "reading-the-routine-study; the-report-somebody-acts-on", "title": "Geoffrey Hinton said in 2016 that people should stop training radiologists now, that within five years deep learning would do better, and added that it might be ten", "stage": "forecast", "occurred_on": "2016-11-24", "verified_on": "2026-09-12", "source_name": "Creative Destruction Lab — Geoff Hinton: On Radiology, recorded at the 2016 Machine Learning and Market for Intelligence Conference, Toronto", "source_url": "https://www.youtube.com/watch?v=2HMPRXstSvQ", "source_tier": "primary", "scope": "The recording of the remark itself, published by the conference organiser, 84 seconds long. Read what he actually said rather than the version that circulates, because they differ in a way that matters: the widely repeated form is five years, and the recording continues with a concession the retellings drop — it might be ten years, but we got plenty of radiologists already. He also framed the claim as a prediction about capability, that deep learning would do better than radiologists because it could accumulate more experience, not as a claim about hospitals, employment or regulation. Two limits on this record itself. The words here were read from the video's auto-generated captions, which are a machine transcription, checked against the audio for the sentences quoted but not professionally transcribed. And a forecast is never evidence about work: this record moves nothing on this page, and sits next to the records of what happened — a regulator's authorisation count and a decade of training-post numbers — which is where a reader can do the arithmetic themselves.", "url": "https://flyvolo.ai/en/changes/ev-20161124-radiologist-4" }, { "id": "ev-20181001-cabin-crew-1", "occupation_slug": "cabin-crew", "task_ids": "the-evacuation; crew-in-charge", "title": "Singapore's air operations rules set the minimum cabin crew by seats, exits and certification, and require a named crew-in-charge answerable to the captain", "stage": "constraint", "occurred_on": "2018-10-01", "verified_on": "2026-09-23", "source_name": "Singapore Statutes Online (Attorney-General's Chambers) — Air Navigation (121 — Commercial Air Transport by Large Aeroplanes) Regulations 2018, regulations 138 and 139, current version", "source_url": "https://sso.agc.gov.sg/SL/ANA1966-S444-2018", "source_tier": "primary", "scope": "The regulations as published on Singapore Statutes Online, in operation since 1 October 2018 and read in the current version. Regulation 138: an operator may not carry passengers on an aeroplane with more than 19 passenger seats unless the cabin crew on board is not less than the greatest of one per 50, or fraction of 50, passenger seats on the same deck; one per pair of directly opposing floor-level exits on a single-aisle aeroplane, or one per floor-level exit with more than one aisle; and the number the manufacturer determined at certification. It allows one crew member fewer only if a crew member is incapacitated, no qualified replacement is reasonably available, seating capacity is reduced, and for one sector unless the regulator approves a second. Regulation 139: every flight needing at least two cabin crew must have a designated crew-in-charge, responsible to the pilot-in-command for the operational and safety functions of each cabin crew member, with at least one year's experience with the operator and appropriate training. It establishes a legal minimum and who is accountable in Singapore; it does not say how many crew airlines roster above that minimum, and it says nothing about onboard service.", "url": "https://flyvolo.ai/en/changes/ev-20181001-cabin-crew-1" }, { "id": "ev-20181201-lab-technician-2", "occupation_slug": "lab-technician", "task_ids": "knowing-the-result-is-wrong", "title": "Japan's medical ordinance requires a hospital doing its own laboratory testing to have a named person responsible for accuracy who must be a physician or a licensed clinical technologist", "stage": "constraint", "occurred_on": "2018-12-01", "verified_on": "2026-09-13", "source_name": "e-Gov 法令検索 (Japan, Digital Agency) — 医療法施行規則 第九条の七・第九条の七の二", "source_url": "https://laws.e-gov.go.jp/law/323M40000100050", "source_tier": "primary", "scope": "Read on the Digital Agency's own statute portal; the requirement entered force on 1 December 2018 under the 2017 amendment to the Medical Care Act and the Clinical Laboratory Technicians Act. Article 9-7 sets out what a facility performing its own specimen testing must have, and the first item is a person responsible for ensuring accuracy — for a medical institution, a physician or a licensed clinical laboratory technologist, and for genetic and chromosomal testing someone with considerable experience of that work specifically. The same article requires standard operating documents for instrument maintenance and for measurement, daily work logs for both, and three ledgers: reagents, statistical quality control, and external quality control. Note the change in verb strength at the next article, because it is the finding. Having the named person is a standard the facility must meet. Article 9-7-2 then says the manager shall endeavour to build a quality-control system centred on that person so internal quality control is carried out, shall endeavour to undergo external quality-control surveys, and shall endeavour to have staff trained. The person is mandatory; the activity that would actually catch a wrong result is best-efforts. This binds facilities in Japan and counts nobody — no tests, no errors caught, no posts.", "url": "https://flyvolo.ai/en/changes/ev-20181201-lab-technician-2" }, { "id": "ev-20190101-customer-service-representative-5", "occupation_slug": "customer-service-representative", "task_ids": "faq-answering", "title": "The FTC listed its 2019 after-hours chatbot for ReportFraud.gov and IdentityTheft.gov as retired in the 2026 federal AI inventory, a year after listing the same use case as in operation", "stage": "deployment", "occurred_on": "2019-01-01", "verified_on": "2026-09-22", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory, individually reported use cases (entry FTC-0003), read against the 2024 inventory", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory/blob/main/Data/2025_individually_reported_AI_use_cases.csv", "source_tier": "primary", "scope": "Both sides of this come from the employer describing its own systems in the inventory it must publish. Entry FTC-0003 in the 2025 file gives an operational date of 2019 with no month, so it is recorded as the first of January, and a development stage of Retired. The same use case appears in the 2024 file at the stage Operation and Maintenance with the Date Retired column empty, and that file was uploaded on 23 January 2025 while the 2025 file is dated as of 13 April 2026 — so the retirement falls between those two dates and nothing narrows it further. Its own wording moved between the two years and the reason for that is not given either: in 2024 the agency described it as delivering unattended chat outside business hours, when assisted chat is unavailable, and in 2025 as providing automated chatbot capability and replacing assisted chat services. The entry names a contractor and says the agency does not hold the source code. Two things this is deliberately not. It is not recorded as a rollback, because the file gives no reason for the retirement and a system can end for a contract, a site rebuild or a better replacement as easily as for failing; recording it under a stage that means suppressed adoption would assert a cause the document does not contain. And it is not a measurement: no volume of chats, no staffing figure, and nothing about what happened to the assisted chat service afterwards. What it does establish is that one employer ran this for years and has stopped, and that the ending is visible only because this employer is legally obliged to keep listing what it uses.", "url": "https://flyvolo.ai/en/changes/ev-20190101-customer-service-representative-5" }, { "id": "ev-20190301-translator-5", "occupation_slug": "translator", "task_ids": "high-stakes", "title": "The US Department of Justice reports machine translation deployed department-wide since March 2019 for an attorney's initial review, and states that its output does not constitute an official record", "stage": "deployment", "occurred_on": "2019-03-01", "verified_on": "2026-09-21", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory (entry DOJ-0244)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory", "source_tier": "primary", "scope": "The entry is the department's own submission to the statutory inventory OMB publishes under the Advancing American AI Act, so the publisher is the employer describing a deployment it runs. The department draws the line itself: the machine output is a tool for initial review by an attorney and, in its own words, does not constitute an official record. The entry gives its operational date only as a month, 03/2019, so it is recorded as the first of that month; the reported date is the April 2026 data upload, which the repository dates as of 13 April 2026. Read across the whole file, the same product appears inside this one department at four bureaus deployed (department-wide from 03/2019, the antitrust division from 01/2023, the criminal division and the public affairs office from 01/2024), one bureau not yet deployed, and one retired. A second agency, the Federal Trade Commission, records the same product deployed since 2022 and gives its reason as manual translation and transcription being slow and of inconsistent accuracy. What the entry does not give is any headcount, any volume, or any measure of how review time changed. It also says nothing about certified translation: it describes an initial read, not an official one, so it does not bear on who signs a certified document.", "url": "https://flyvolo.ai/en/changes/ev-20190301-translator-5" }, { "id": "ev-20190901-it-support-specialist-3", "occupation_slug": "it-support-specialist", "task_ids": "the-repeat-ticket", "title": "The US Department of Justice reports a department-wide system, operational since September 2019, that triages and classifies IT helpdesk tickets and handles common requests with virtual agents", "stage": "deployment", "occurred_on": "2019-09-01", "verified_on": "2026-09-21", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory (entry DOJ-0107)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory", "source_tier": "primary", "scope": "The entry is the department's own submission to the statutory inventory OMB publishes under the Advancing American AI Act, so the publisher here is the employer describing a deployment it runs, not a vendor or a journalist. The entry gives its operational date only as a month, 09/2019, so it is recorded as the first of that month; the reported date is the April 2026 data upload, which the repository README dates as of 13 April 2026. Read across the whole 2025 file and hand-filtered to genuine IT service-desk use cases, the same inventory holds 12 deployed across 6 agencies, 4 pilots, 26 in pre-deployment across 11 agencies and 1 retired — more are waiting than are running. What the entry does not give is any headcount figure, and it records no measurement of what share of the queue the system closes. Because this deployment predates public generative models by three years, it is evidence about ticket triage automation, not about language models.", "url": "https://flyvolo.ai/en/changes/ev-20190901-it-support-specialist-3" }, { "id": "ev-20200204-loan-officer-2", "occupation_slug": "loan-officer", "task_ids": "scoring-the-application", "title": "Korea's Credit Information Act gives an individual the right to be told the main criteria of an automated credit evaluation and to demand the result be recalculated", "stage": "constraint", "occurred_on": "2020-02-04", "verified_on": "2026-09-13", "source_name": "국가법령정보센터 (Korea Ministry of Government Legislation) — 신용정보의 이용 및 보호에 관한 법률 제36조의2", "source_url": "https://www.law.go.kr/법령/신용정보의이용및보호에관한법률/제36조의2", "source_tier": "primary", "scope": "The statute read article by article on the Ministry of Government Legislation's own portal, in the version in force from 11 September 2026; the article itself was newly inserted on 4 February 2020, which is the date recorded here. Paragraph 1 lets an individual require a credit rating company, or a prescribed credit information provider or user, to say whether automated evaluation is used for personal credit rating and for the setting, maintenance and terms of prescribed financial transactions — and if it is, to disclose the result, its main criteria, and an outline of the base information used. Paragraph 2 goes further than disclosure: the person may submit information they consider favourable, and where the base information is inaccurate or out of date may require it to be corrected or deleted and the automated result to be recalculated. Read paragraph 3 with the rest or the record misstates the law: the firm may refuse any of this where another law requires it, where refusal is unavoidable to meet a legal duty, or where complying would make it difficult to establish or maintain the commercial relationship — the last of which is broad and is the kind of clause a retelling drops. Procedure is left to Presidential Decree. This binds firms operating in Korea and says nothing about lending anywhere else. What it establishes about work is indirect but real: a legislature wrote a recourse route on the assumption that this scoring is already done by machine.", "url": "https://flyvolo.ai/en/changes/ev-20200204-loan-officer-2" }, { "id": "ev-20200401-auto-mechanic-2", "occupation_slug": "auto-mechanic", "task_ids": "reading-the-fault; doing-the-repair", "title": "Japan's transport ministry made camera and radar calibration a licensed category of work needing certification, a trained supervisor and a scan tool", "stage": "mandate", "occurred_on": "2020-04-01", "verified_on": "2026-09-13", "source_name": "国土交通省自動車局——《自動車特定整備事業について》", "source_url": "https://www.mlit.go.jp/jidosha/jidosha_fr9_000016.html", "source_tier": "primary", "scope": "The ministry's own page on the certified-maintenance regime. From April 2020 the old category of disassembly maintenance was widened and renamed specific maintenance, and two things were added. The first is a new kind of work: maintenance or modification that affects how a device operates even without removing it — named concretely as adjusting the forward-facing cameras and radar used by automatic braking. The second is a new target device: the automated driving system fitted to vehicles at level 3 and above. A garage doing either needs certification from the regional transport bureau, a dedicated electronic-control inspection and maintenance bay, diagnostic scan tools, and at least one maintenance supervisor who has completed the ministry's qualification course — first-class small-vehicle mechanics are exempt from that course. Read the direction carefully, because it runs the opposite way from most records on this site: driving automation did not remove work here, it created a licensed category of it and put a training requirement and a tool purchase in front of anyone who wants it. What this does not establish is how many garages obtained the certification, what it cost them, or whether independent shops kept pace with dealers. It applies in Japan.", "url": "https://flyvolo.ai/en/changes/ev-20200401-auto-mechanic-2" }, { "id": "ev-20210511-government-service-clerk-2", "occupation_slug": "government-service-clerk", "task_ids": "taking-the-application", "title": "Singapore's communications ministry told Parliament on 11 May 2021 that the Digital Government Blueprint target of 90–95% of government transactions completed digitally end to end had been met", "stage": "deployment", "occurred_on": "2021-05-11", "verified_on": "2026-09-12", "source_name": "Ministry of Digital Development and Information (then MCI), Singapore — written answer to Parliamentary Question, 11 May 2021", "source_url": "https://www.mddi.gov.sg/newsroom/mci-response-pq-complete-paperless-processes-transactions/", "source_tier": "primary", "scope": "Singapore only, and it is the ministry's own assertion that a target was met — the answer does not publish what counts as a transaction, which agencies are in scope, or how it was measured, so the figure is a policy statement rather than an audited statistic. Note the date above all: May 2021, before generative AI existed as a public tool. What moved the routine counter transaction online in this market was ordinary digital government — forms, national identity, pre-filled data — and it had largely happened already. That matters for how the rest of this page reads: the open question for this occupation is not whether the routine application leaves the counter, but what happens to the cases that could not.", "url": "https://flyvolo.ai/en/changes/ev-20210511-government-service-clerk-2" }, { "id": "ev-20210621-registered-nurse-4", "occupation_slug": "registered-nurse", "task_ids": "monitoring-and-escalation", "title": "An external validation of the Epic Sepsis Model at Michigan Medicine found an AUC of 0.63, missing 67% of sepsis patients while alerting on 18% of all hospitalised patients", "stage": "constraint", "occurred_on": "2021-06-21", "verified_on": "2026-09-11", "source_name": "JAMA Internal Medicine — Wong et al. 181(8) (open-access copy, PMC8218233)", "source_url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC8218233/", "source_tier": "primary", "scope": "27,697 patients over 38,455 hospitalisations at one US academic health system, 6 December 2018 to 20 October 2019; sepsis in 6.6% of hospitalisations. At the implemented threshold the model found 183 of 2,552 sepsis patients (7%) that clinical care had missed. The vendor's own reported AUC was 0.76-0.83. One widely deployed proprietary model at one site — not every early-warning tool, and Epic revised the model afterwards.", "url": "https://flyvolo.ai/en/changes/ev-20210621-registered-nurse-4" }, { "id": "ev-20211102-real-estate-agent-1", "occupation_slug": "real-estate-agent", "task_ids": "pricing-the-property", "title": "Zillow wound down Zillow Offers, its algorithmic home-buying business, cutting about 25% of staff; its CEO wrote that unpredictability in forecasting home prices far exceeded expectations", "stage": "constraint", "occurred_on": "2021-11-02", "verified_on": "2026-09-12", "source_name": "Zillow Group (Exhibit 99.1 to Form 8-K, SEC EDGAR)", "source_url": "https://www.sec.gov/Archives/edgar/data/1617640/000161764021000085/q32021991.htm", "source_tier": "primary", "scope": "One US company taking balance-sheet risk on its own automated valuations, during an unusually volatile housing market and alongside stated capacity and supply-chain constraints. It says nothing about the accuracy of automated valuation as a tool, and nothing about whether agents price better — the company kept publishing estimates after closing the buying business. What it establishes is about who carries the consequence of a valuation, not who produces one.", "url": "https://flyvolo.ai/en/changes/ev-20211102-real-estate-agent-1" }, { "id": "ev-20211115-firefighter-2", "occupation_slug": "firefighter", "task_ids": "the-attack", "title": "SCDF's robotics staff officer forecast nearly 50% fewer firefighters committed into high-risk areas once its two-tier robot approach was operational", "stage": "forecast", "occurred_on": "2021-11-15", "verified_on": "2026-09-23", "source_name": "Singapore Civil Defence Force — Rescue 995, Leveraging Robotics in SCDF (interview)", "source_url": "https://www.scdf.gov.sg/home/about-scdf/media-room/our-stories/storyarticledetail/leveraging-robotics-in-scdf", "source_tier": "primary", "scope": "An interview published by SCDF itself on 15 November 2021 in its in-house magazine. The officer describes a two-tier approach — small robots at every fire station for small and medium fires, heavier machines placed at certain stations for large fires — and says: with this tiered approach, we forecast a nearly 50% reduction in the number of firefighters needed to be committed into the high-risk areas during offensive firefighting operations, with the approach to be fully operationalised by 2023. He also states that the heavy Unmanned Firefighting Machine is remotely controlled by one person, that robots are deployed manually from vehicles on arrival, and that autonomous deployment and swarm coordination are later stages of a ten-year roadmap. The number is about how many people go into the most dangerous zone during an attack, not about how many firefighters the service employs. A forecast is not evidence about work; this record moves nothing on this page and sits beside the service's own later report of what was deployed.", "url": "https://flyvolo.ai/en/changes/ev-20211115-firefighter-2" }, { "id": "ev-20211231-farmer-3", "occupation_slug": "farmer", "task_ids": "the-livestock", "title": "USDA's Economic Research Service reported that robotic milking produced 6% of U.S. milk in 2021, up from 4% in 2016, with adoption highest on midsized dairy farms", "stage": "deployment", "occurred_on": "2021-12-31", "verified_on": "2026-09-23", "source_name": "USDA Economic Research Service — Robotic milking gains ground, especially among midsized dairies (Chart of Note, citing ERR-356)", "source_url": "https://ers.usda.gov/data-products/charts-of-note/114194", "source_tier": "primary", "scope": "USDA's Economic Research Service (ERS), citing its own Economic Research Report ERR-356 (Precision Dairy Farming, Robotic Milking, and Profitability in the United States, published 22 January 2026, using the department's Agricultural Resource Management Survey), reported that robotic milking systems produced 6 percent of U.S. milk in 2021, up from 4 percent in 2016. Adoption is uneven by farm size: in 2016 it was higher on larger farms, but by 2021 the pattern had shifted so that 13 percent of dairy farms with 150 to 499 head used robotic milking, the highest of any size class. ERS attributes the shift to economics rather than the technology's limits: farms under 50 cows rely on unpaid family labor, so the capital cost is harder to justify, and large farms already have low per-unit labor costs and would need to rebuild their milking parlours to add robots. This is United States dairy only, describes production volume and one survey year (2021), and says nothing about adoption after that year or in other countries. It does not say how many individual farmers operate the systems themselves versus hire the work out.", "url": "https://flyvolo.ai/en/changes/ev-20211231-farmer-3" }, { "id": "ev-20220510-dentist-2", "occupation_slug": "dentist", "task_ids": "reading-xrays", "title": "The FDA cleared AI software that marks suspected tooth decay on bitewing X-rays, on the condition that trained dentists use it and make the diagnosis", "stage": "capability", "occurred_on": "2022-05-10", "verified_on": "2026-09-24", "source_name": "US Food and Drug Administration — 510(k) K212519, Overjet Caries Assist (clearance letter and summary)", "source_url": "https://www.accessdata.fda.gov/cdrh_docs/pdf21/K212519.pdf", "source_tier": "primary", "scope": "The regulator's clearance letter of 10 May 2022 with the device's indications for use and 510(k) summary, read in the full PDF. The software is cleared to aid in detecting and outlining caries on bitewing radiographs, providing additional information for the dentist; it is not intended as a replacement for a complete dentist's review or clinical judgement, should be used only by trained dentists, assists only in detection and not interpretation or diagnosis, and should not be the sole decision-making tool for diagnosis or treatment; final clinical determination is the treating clinician's. The summary reports the manufacturer's own testing: standalone sensitivity 72.0% and specificity 98.1% on 352 images; in a reader study with 13 US-licensed dentists, sensitivity rose from 57.9% unaided to 76.2% aided while specificity fell from 99.3% to 98.4%. It lists limits including decay hidden by restorations, gross decay and endodontic access mistaken for caries. It establishes what one tool may be sold to do and its manufacturer's reported performance; it is not a measure of use in practice, and a capability record moves nothing on this page.", "url": "https://flyvolo.ai/en/changes/ev-20220510-dentist-2" }, { "id": "ev-20220601-firefighter-1", "occupation_slug": "firefighter", "task_ids": "the-attack; the-long-fire", "title": "Singapore's Civil Defence Force moved to three-person fire sections and put small firefighting robots on its new fire engines", "stage": "deployment", "occurred_on": "2022-06-01", "verified_on": "2026-09-23", "source_name": "Singapore Civil Defence Force — Annual Report FY 2022–2023", "source_url": "https://www.scdf.gov.sg/docs/default-source/media-room-(publications)/annual-reports/annual-report-22-23_compressed.pdf?sfvrsn=8cdb1155_1", "source_tier": "primary", "scope": "The service's own annual report for FY 2022–2023, read in the full PDF. It states that SCDF transitioned to a three-crew section concept of operations from 1 June 2022 and launched the sixth-generation Light Fire Attack Vehicle and the Red Rhino Robot at that point; that the robot and vehicle were designed to complement the three-crew concept, in a project described as using human-factors technologies to alleviate manpower constraints; that Tier 1 ground robots — the Red Rhino Robot and the Pumper Firefighting Machine — are readily available in all new fire engines for immediate deployment at any fire scene; and, in its incident list, that an Unmanned Firefighting Machine was deployed at warehouse and industrial fires on 18 June 2022, 17 September 2022 and 13 March 2023, alongside about 50 and about 70 firefighters at the two fires where a number is given. It establishes a deployment and a change in crew size at one national service, reported together; it does not say the robots caused the smaller sections, how many firefighters the service employs, or how often the robots are used.", "url": "https://flyvolo.ai/en/changes/ev-20220601-firefighter-1" }, { "id": "ev-20220928-government-service-clerk-1", "occupation_slug": "government-service-clerk", "task_ids": "taking-the-application", "title": "China's State Council General Office required cross-provincial government services to run through the national integrated platform, with listed items standardised nationwide by mid-2023", "stage": "mandate", "occurred_on": "2022-09-28", "verified_on": "2026-09-12", "source_name": "国务院办公厅 · 国办发〔2022〕34号(中国政府网)", "source_url": "https://www.gov.cn/zhengce/content/2022-10/05/content_5715850.htm", "source_tier": "primary", "scope": "Mainland China, issued by the State Council General Office. It requires adoption and sets a deadline for standardising the names, codes and legal bases of listed cross-provincial items; it also pushes mutual recognition of commonly used electronic certificates and adds 22 items in its annex. It is a requirement, not a report of what happened: it says nothing about how many counters closed, what share of applicants complete online, or staffing. The same document requires online and offline channels to be deeply integrated rather than counters closing.", "url": "https://flyvolo.ai/en/changes/ev-20220928-government-service-clerk-1" }, { "id": "ev-20221001-cybersecurity-analyst-2", "occupation_slug": "cybersecurity-analyst", "task_ids": "alert-triage", "title": "Singapore's infocomm regulator assessed two security-analyst roles as high-impact, expecting SOAR tooling to replace their core tasks within three to five years", "stage": "forecast", "occurred_on": "2022-10-01", "verified_on": "2026-09-13", "source_name": "IMDA / Workforce Singapore — Impact Study on the Information & Communications Workforce in Singapore (full report, PDF)", "source_url": "https://www.swda.gov.sg/api/assets/d65e963c-df80-4768-b5b4-4bb214307a22/imda-impact-study-master-full-vf6.pdf", "source_tier": "primary", "scope": "A commissioned sector study, not a measurement of anything that happened, which is why it sits in a stage that changes nothing. Read the mechanism carefully, because it is not the one most readers will assume: the tool named is SOAR — security orchestration, automation and response — which is workflow automation, not a language model. The report is dated October 2022 and its underlying analysis is dated December 2020, so it was written before generative AI reached the public; 'AI and analytics' appears in it as a trend heading covering machine learning for alert prioritisation, not chat models. What it claims specifically is that SOAR will take over manual cyber monitoring and reporting and potentially replace core tasks such as managing cyber security systems and operations, thereby reducing the manpower required for those tasks, while job holders keep the analysis of log data and reports. It covers Singapore's infocomm workforce only, it aggregates stakeholder interviews rather than measuring employment, and the report itself says its findings were taken at a point in time and must be re-contextualised by whoever reads them later.", "url": "https://flyvolo.ai/en/changes/ev-20221001-cybersecurity-analyst-2" }, { "id": "ev-20221001-devops-engineer-2", "occupation_slug": "devops-engineer", "task_ids": "the-pager", "title": "Singapore's infocomm regulator assessed that service-level and system-performance duties would move onto DevOps teams, and named automation and orchestration engineers among the growing roles", "stage": "forecast", "occurred_on": "2022-10-01", "verified_on": "2026-09-13", "source_name": "IMDA / Workforce Singapore — Impact Study on the Information & Communications Workforce in Singapore (full report, PDF)", "source_url": "https://www.swda.gov.sg/api/assets/d65e963c-df80-4768-b5b4-4bb214307a22/imda-impact-study-master-full-vf6.pdf", "source_tier": "primary", "scope": "A commissioned sector study, not a measurement, and it is recorded here because it points the other way from most of this base. The same document that expects stand-alone support engineering to shrink expects the work to land on this role: it says the DevOps function will take over overseeing service level agreements and developing new systems, that development and operations come together to provide more holistic support, and it lists automation and orchestration engineers among the roles growing in demand. It also names DevOps Engineer as the destination it considers an easy or moderate move for the support engineers it expects to be displaced. None of that is evidence that anyone was hired, that headcount rose, or that the transition happened — it is one agency's expectation, published before generative AI reached the public, covering Singapore only and built from stakeholder interviews rather than employment data. What it does establish is that the displacement finding elsewhere in the same report is not a claim that the work disappears.", "url": "https://flyvolo.ai/en/changes/ev-20221001-devops-engineer-2" }, { "id": "ev-20221001-it-support-specialist-2", "occupation_slug": "it-support-specialist", "task_ids": "the-repeat-ticket", "title": "Singapore's infocomm regulator assessed stand-alone applications and systems support engineer roles as high-impact, judging their tasks would be absorbed into development and DevOps teams", "stage": "forecast", "occurred_on": "2022-10-01", "verified_on": "2026-09-13", "source_name": "IMDA / Workforce Singapore — Impact Study on the Information & Communications Workforce in Singapore (full report, PDF)", "source_url": "https://www.swda.gov.sg/api/assets/d65e963c-df80-4768-b5b4-4bb214307a22/imda-impact-study-master-full-vf6.pdf", "source_tier": "primary", "scope": "A commissioned sector study, not a measurement, and the mechanism it names is the part worth keeping. Only one of its three routes is a machine doing the work: AI automating help desk answers and ticket routing. The other two are reorganisation — under Agile and CI/CD the development function takes over maintaining software and overseeing transition and testing, and under SaaS or PaaS the service provider takes over maintaining and updating the software, 'reducing the manpower needed for the in-house capability'. The report's own sentence is that there will be a decreasing need for stand-alone roles because the tasks will be 'subsumed under the DevOps team's job tasks'. Work moving to a different team is not the same event as work being automated, and a retelling flattens exactly that difference. Dated October 2022 on analysis dated December 2020, so written before generative AI reached the public — which matters because the help-desk automation it describes is ticket routing and canned answers, not a chat model. Singapore's infocomm workforce only; stakeholder interviews, not employment data.", "url": "https://flyvolo.ai/en/changes/ev-20221001-it-support-specialist-2" }, { "id": "ev-20221221-waiter-2", "occupation_slug": "waiter", "task_ids": "carrying; running-the-room", "title": "Japan's largest family-restaurant group announced it had put 3,000 floor-service robots into about 2,100 of its roughly 2,980 domestic restaurants", "stage": "deployment", "occurred_on": "2022-12-21", "verified_on": "2026-09-13", "source_name": "株式会社すかいらーくホールディングス(東証プライム 3197)——《約2,100店に3,000台のロボット導入完了》", "source_url": "https://corp.skylark.co.jp/Portals/0/images/news/press_release/2022/20221221_Robot.pdf", "source_tier": "primary", "scope": "The operator's own release, from a company listed on the Tokyo Stock Exchange Prime market. It states that floor-service robot deployment began in August 2021 and that 3,000 units would be in place across about 2,100 stores by 27 December 2022 — the release is dated the 21st, so the completion date it names is the company's own announced schedule rather than a fact the document establishes. The store breakdown is given: Gusto 1,251, Bamiyan 354, Shabu-yo 271, Jonathan 187, Uoyaro 14, Steak Gusto 3, Yumean 2, against a stated domestic group total of 2,982 stores. The reason this belongs on this page is the effects the company chose to report, and the ones it did not. For Gusto it names improved table turnover at the lunch peak, reduced time to finish clearing a table, and fewer steps walked, alongside a customer survey at 67 stores with 2,513 responses in which nine in ten chose satisfied or very satisfied. It reports no headcount, no hours, and no change to staffing of any kind. It also records that the company sent dedicated robot instructors to every store and iterated on staff and customer feedback, which is a cost of deployment that a shorter account leaves out. One operator, one country, one format — table-service family restaurants.", "url": "https://flyvolo.ai/en/changes/ev-20221221-waiter-2" }, { "id": "ev-20230622-lawyer-1", "occupation_slug": "lawyer", "task_ids": "supervising-machine-work", "title": "A US federal court fined two attorneys and their firm $5,000 for a brief citing six non-existent cases generated by ChatGPT, noting that using a reliable AI tool is not itself improper", "stage": "constraint", "occurred_on": "2023-06-22", "verified_on": "2026-09-10", "source_name": "Mata v. Avianca, Inc. (S.D.N.Y., 22 June 2023) — opinion and order on sanctions, via CourtListener", "source_url": "https://www.courtlistener.com/docket/63107798/54/mata-v-avianca-inc/", "source_tier": "primary", "scope": "US federal court (S.D.N.Y.). Establishes that the signing lawyer, not the tool, carries responsibility for machine-assisted filings; several bars issued guidance afterwards.", "url": "https://flyvolo.ai/en/changes/ev-20230622-lawyer-1" }, { "id": "ev-20230622-paralegal-1", "occupation_slug": "paralegal", "task_ids": "verification-of-machine-output", "title": "A US federal court fined two attorneys and their firm $5,000 for a brief citing six non-existent cases generated by ChatGPT, noting that using a reliable AI tool is not itself improper", "stage": "constraint", "occurred_on": "2023-06-22", "verified_on": "2026-09-10", "source_name": "Mata v. Avianca, Inc. (S.D.N.Y., 22 June 2023) — opinion and order on sanctions, via CourtListener", "source_url": "https://www.courtlistener.com/docket/63107798/54/mata-v-avianca-inc/", "source_tier": "primary", "scope": "US federal court (S.D.N.Y., Judge Castel). The sanction rested on failing to verify and on standing by the citations once challenged. Widely cited in later professional-conduct guidance.", "url": "https://flyvolo.ai/en/changes/ev-20230622-paralegal-1" }, { "id": "ev-20231110-retail-cashier-1", "occupation_slug": "retail-cashier", "task_ids": "checkout", "title": "UK supermarket chain Booths removed self-checkouts from 26 of its 28 stores and returned to staffed tills, citing machines that were slow, unreliable and impersonal and trouble with loose produce", "stage": "constraint", "occurred_on": "2023-11-10", "verified_on": "2026-09-10", "source_name": "Fortune (quoting BBC)", "source_url": "https://fortune.com/2023/11/10/grocery-chain-removes-self-checkout-we-like-to-talk-to-people/", "source_tier": "secondary", "scope": "One regional UK chain (28 stores); a reversal, not the industry norm. Some larger chains later trimmed self-checkout too; others kept expanding it.", "url": "https://flyvolo.ai/en/changes/ev-20231110-retail-cashier-1" }, { "id": "ev-20231121-ride-hail-driver-4", "occupation_slug": "ride-hail-driver", "task_ids": "fleet-support-roles; the-drive", "title": "China's transport ministry requires an onboard safety officer in conditionally and highly automated taxis, and caps remote safety officers at one per three fully driverless taxis", "stage": "constraint", "occurred_on": "2023-11-21", "verified_on": "2026-09-12", "source_name": "中华人民共和国交通运输部办公厅 —《自动驾驶汽车运输安全服务指南(试行)》,交办运〔2023〕66 号", "source_url": "https://xxgk.mot.gov.cn/2020/jigou/ysfws/202312/t20231205_3962490.html", "source_tier": "primary", "scope": "Mainland China, and the document is a trial guideline from the ministry's General Office rather than a statute, so its force runs through provincial transport authorities and the licences they issue. Read it as two things at once: it legitimises commercial autonomous passenger and freight operation, and in the same breath it sets a personnel floor. Section 6 requires one onboard safety officer in a conditionally or highly automated taxi; a fully automated taxi may use a remote officer only in a designated area, with the municipal government's consent, and at a ratio no lower than one officer per three vehicles. Officers must be trained on the specific automation level and route and must hold the relevant professional qualification. The ratio is a floor on that permitted configuration, not a measurement of what operators actually staff, and it says nothing about any other market — nor about how it will be enforced.", "url": "https://flyvolo.ai/en/changes/ev-20231121-ride-hail-driver-4" }, { "id": "ev-20231121-truck-driver-4", "occupation_slug": "truck-driver", "task_ids": "remote-supervision; load-responsibility", "title": "China's transport ministry says autonomous road freight should in principle carry an onboard safety officer, and bans autonomous vehicles from carrying dangerous goods", "stage": "constraint", "occurred_on": "2023-11-21", "verified_on": "2026-09-12", "source_name": "中华人民共和国交通运输部办公厅 —《自动驾驶汽车运输安全服务指南(试行)》,交办运〔2023〕66 号", "source_url": "https://xxgk.mot.gov.cn/2020/jigou/ysfws/202312/t20231205_3962490.html", "source_tier": "primary", "scope": "Mainland China, and a trial guideline from the ministry's General Office rather than a statute — its force runs through provincial authorities and operating permits. Two clauses matter for freight and they pull in opposite directions from each other. Section 3 permits autonomous freight on point-to-point trunk highways and on urban roads where safety is controllable, which is a green light; it then bans autonomous vehicles outright from dangerous-goods haulage. Section 6 says autonomous road freight should in principle carry an onboard safety officer — note in principle, which is markedly weaker than the taxi and bus clauses in the same section, where an officer is simply required. The load-responsibility duties elsewhere in the guideline sit on the operator, not on a person in the cab. This describes what is permitted, not what is running: it is not evidence that any Chinese carrier has put a driverless truck into service.", "url": "https://flyvolo.ai/en/changes/ev-20231121-truck-driver-4" }, { "id": "ev-20231210-bus-driver-4", "occupation_slug": "bus-driver", "task_ids": "safety-operator; getting-people-on", "title": "A four-month evaluation of the Treasure Island \"Loop\" autonomous shuttle pilot in San Francisco recorded its on-board attendant taking manual control 358 times and deploying a wheelchair ramp 18 times", "stage": "pilot", "occurred_on": "2023-12-10", "verified_on": "2026-09-24", "source_name": "Treasure Island Mobility Management Agency — The Loop Final Evaluation Report", "source_url": "https://www.sfcta.org/sites/default/files/2024-06/Loop_Final_Evaluation_Report_2024-06-25_0.pdf", "source_tier": "primary", "scope": "The Treasure Island Mobility Management Agency's own final evaluation report (a joint powers agency formed by the San Francisco County Transportation Authority, the Treasure Island Development Authority and the San Francisco Municipal Transportation Agency), funded by a USDOT/FHWA Smart City grant, an MTC grant and local Prop K sales-tax funds. It covers one closed pilot on Treasure Island, San Francisco: three Level-3 shuttles operated by Beep with GMM (formerly Navya) vehicles on a fixed route from 16 August to 10 December 2023, each carrying an on-board attendant/safety driver at all times. The report's own operations and accessibility sections give the attendant's tasks in counts, not a plan: 358 disengagements over the four months, where the vehicle handed control to the attendant, most often because another road user or an object was detected in the shuttle's path, or because the 5G/GNSS/RTK signal was lost; and 18 ADA ramp deployments plus 5 wheelchair securements to board passengers. occurredOn is dated to the pilot's last day of revenue service (10 December 2023), since every count above is a total across the whole evaluated period; reportedOn is the final report's own publication, dated June 2024 on its cover and 25 June 2024 in the file's own path. This is a different country, operator and regulator (the US National Highway Traffic Safety Administration and the California DMV, not Singapore's Land Transport Authority or Japan's transport bureau) from the two pilots already on this page, so it stands as an independent record rather than a restatement of them. It also cuts against reading this cleanly: in one incident the shuttle detected another vehicle running a stop sign and braked, but the attendant did not engage the emergency-stop button before a low-speed, no-injury collision. The pilot ended early, on 10 December 2023, because remapping the route around changed road conditions was cost-prohibitive for the agency to fund again — the report frames this as an operational and funding limit, not as a finding that the attendant role could be removed.", "url": "https://flyvolo.ai/en/changes/ev-20231210-bus-driver-4" }, { "id": "ev-20231218-architect-1", "occupation_slug": "architect", "task_ids": "documentation-and-detailing; approvals-and-coordination", "title": "Singapore's BCA soft-launched CORENET X on 18 Dec 2023: one coordinated BIM model replaces separate agency submissions, ~20 approval stages become 3 gateways; automated geometric checks in development", "stage": "deployment", "occurred_on": "2023-12-18", "verified_on": "2026-09-10", "source_name": "CORENET X (Singapore Building and Construction Authority) — overview", "source_url": "https://info.corenet.gov.sg/overview/about-corenet-x/overview-of-corenet-x", "source_tier": "primary", "scope": "Singapore regulatory submissions only; mandatory for new projects of 30,000 m² GFA or more from 1 Oct 2025 (info.corenet.gov.sg). Automated checks cover straightforward geometric and spatial rules across architecture, C&S and M&E; judgement-based compliance stays with the qualified person.", "url": "https://flyvolo.ai/en/changes/ev-20231218-architect-1" }, { "id": "ev-20231231-insurance-claims-handler-1", "occupation_slug": "insurance-claims-handler", "task_ids": "taking-the-claim", "title": "Lemonade's annual report states that as of 31 December 2023, 98% of the time its claims bot takes the first notice of loss and pays or declines the claim without human intervention", "stage": "deployment", "occurred_on": "2023-12-31", "verified_on": "2026-09-12", "source_name": "Lemonade, Inc. (Form 10-K for FY2023, SEC EDGAR)", "source_url": "https://www.sec.gov/Archives/edgar/data/1691421/000169142124000026/lmnd-20231231.htm", "source_tier": "primary", "scope": "One US direct-to-consumer insurer writing simple personal lines on systems it built itself. The filing's very next sentence limits the claim: claims the bot is not authorised to settle, or where it identifies concerns, are triaged and assigned to human claims experts. So this establishes automation of the intake step at this company, not that claims work no longer needs people, and it says nothing about traditional insurers running older policy systems — which hold most premium in most markets.", "url": "https://flyvolo.ai/en/changes/ev-20231231-insurance-claims-handler-1" }, { "id": "ev-20231231-train-driver-1", "occupation_slug": "train-driver", "task_ids": "driving-between-stations", "title": "UITP counted 2,279 km of automated metro lines worldwide at the end of 2023, which is 11% of all metro kilometres, and sixty urban areas with at least one such line against thirteen in 2000", "stage": "deployment", "occurred_on": "2023-12-31", "verified_on": "2026-09-20", "source_name": "UITP — Global Metro Figures 2024 (Statistics Brief, May 2025; data as of 31 December 2023)", "source_url": "https://www.uitp.org/wp-content/uploads/sites/7/2025/08/20250822_Global-Metro-Figures_Statistics-Brief_WEB.pdf", "source_tier": "primary", "scope": "Worldwide, and it counts kilometres rather than people. The publisher is an international association of public transport operators and suppliers, reporting through its own Observatory of Automated Metros, which in its own words gathers the world's leading operators with experience in full automated metro operation — so this is an industry body counting its own members' systems, and the same series carries a section headed the case for automation. The definition is the load-bearing part and it is narrow: automated here means designed for operation without staff on board of the trains, with a defining characteristic being the absence of a driver's cabin, which the standard IEC 62267 calls Grade of Automation 4. Every line where the train accelerates and brakes itself while a driver still closes the doors is therefore absent from these figures, so the 11% is the share with nobody aboard and not the share where driving is automatic. What it does not establish: no staffing number, no ratio of control-room staff to trains, nothing about what happened to anyone. It also cannot be compared line for line against the same publisher's earlier briefs, which state that historical figures are revised. One thing it does settle: the same publisher forecast in its 2019 brief that a further 2,000 km would be commissioned within five years, tripling the total; its own later series shows the stock going from 1,133 km to 2,279 km over that window, which is roughly 57% of the forecast increment.", "url": "https://flyvolo.ai/en/changes/ev-20231231-train-driver-1" }, { "id": "ev-20240124-cybersecurity-analyst-1", "occupation_slug": "cybersecurity-analyst", "task_ids": "alert-triage; hunting", "title": "Britain's national cyber authority assessed that AI will almost certainly increase the volume of cyber attacks over two years, with the uplift concentrated in reconnaissance and social engineering", "stage": "forecast", "occurred_on": "2024-01-24", "verified_on": "2026-09-13", "source_name": "UK National Cyber Security Centre, assessment: the near-term impact of AI on the cyber threat", "source_url": "https://www.ncsc.gov.uk/report/impact-of-ai-on-cyber-threat", "source_tier": "primary", "scope": "An intelligence-style assessment using calibrated probability language, not a measurement — 'almost certainly' and 'highly likely' are defined terms on the NCSC's own scale, which the report publishes alongside the judgements. Its relevance to this occupation is indirect but specific: more attacks and more effective social engineering mean a larger and noisier alert queue, which is the most exposed task on this page, so the forecast and the automation point at the same place from opposite directions. The assessment also judges that the near-term uplift comes from evolution of existing techniques rather than novel ones, and that sophisticated uses stay restricted to actors with quality training data, expertise and resources — both of which cut against the most dramatic readings of it.", "url": "https://flyvolo.ai/en/changes/ev-20240124-cybersecurity-analyst-1" }, { "id": "ev-20240214-software-tester-1", "occupation_slug": "software-tester", "task_ids": "test-case-writing", "title": "Meta reported deploying TestGen-LLM at Instagram and Facebook test-a-thons: it improved 11.5% of the classes it was applied to and engineers accepted 73% of its recommended test cases into production", "stage": "pilot", "occurred_on": "2024-02-14", "verified_on": "2026-09-10", "source_name": "Meta — industry paper (arXiv 2402.09171, FSE 2024)", "source_url": "https://arxiv.org/abs/2402.09171", "source_tier": "primary", "scope": "One company, improvement of existing unit-test classes with build/pass/coverage filters; 75% of generated cases built, 57% passed reliably, 25% added coverage. Company-authored paper; a time-boxed test-a-thon setting rather than routine pipeline use.", "url": "https://flyvolo.ai/en/changes/ev-20240214-software-tester-1" }, { "id": "ev-20240227-customer-service-representative-1", "occupation_slug": "customer-service-representative", "task_ids": "faq-answering; triage", "title": "Klarna's AI assistant handled 2.3 million conversations — two-thirds of its customer-service chats — in its first month; resolution time fell from 11 to under 2 minutes", "stage": "deployment", "occurred_on": "2024-02-27", "verified_on": "2026-09-10", "source_name": "Klarna — press release", "source_url": "https://www.klarna.com/international/press/klarna-ai-assistant-handles-two-thirds-of-customer-service-chats-in-its-first-month/", "source_tier": "primary", "scope": "Global consumer fintech, chat channel, 23 markets; refunds, returns, payments, disputes. Company-reported figures; customers could still choose a human agent.", "url": "https://flyvolo.ai/en/changes/ev-20240227-customer-service-representative-1" }, { "id": "ev-20240301-train-driver-3", "occupation_slug": "train-driver", "task_ids": "taking-it-back-by-hand; driving-between-stations", "title": "Japan's transport ministry reported that on JR Kyushu's Kashii Line about 40% of trains are now run by conductors without a driving licence, trained in about two months against about nine for a driver", "stage": "deployment", "occurred_on": "2024-03-01", "verified_on": "2026-09-20", "source_name": "国土交通白書2025 — コラム:鉄道の自動運転(GOA2.5)運転士による乗務が不要に(九州旅客鉄道)", "source_url": "https://www.mlit.go.jp/hakusyo/mlit/r06/hakusho/r07/html/n1212c05.html", "source_tier": "primary", "scope": "One line, one operator, one country, and it is conventional railway rather than metro — the line has level crossings and no platform screen doors, which is what makes it the boundary case rather than the ordinary one. The ministry states the service began in March 2024 as the first automatic operation in Japan on a commercial line with level crossings and without platform doors, and that as of the March 2025 timetable revision about 40% of its trains are run by conductors who do not hold a driving licence. The split of tasks is stated precisely and is the useful part: speed regulation between stations is entirely automatic, while the attendant performs the start request and the door operation, and presses the emergency stop button on noticing an abnormality. Training is the number worth carrying: the operator says it normally takes about nine months to make a driver from a conductor, and about two months to make one of these attendants, because no manual driving proficiency has to be trained; what is trained instead is emergency stopping and abnormal-condition response, and the attendant works alone. The ministry also records the operator expanding this only after first securing drivers able to drive manually in abnormal conditions, and gives the stated motive as a shrinking labour supply rather than cost. What it does not establish: no headcount, no redundancy, and no claim that anyone lost a job — a licensed driver becomes a reserve rather than a rostered post, and whether that is fewer people in total is not stated. It is also a grade the international standard does not define, invented domestically to put a person at the front who is not a driver.", "url": "https://flyvolo.ai/en/changes/ev-20240301-train-driver-3" }, { "id": "ev-20240401-care-worker-2", "occupation_slug": "care-worker", "task_ids": "being-the-company; noticing-the-change", "title": "Japan's health ministry cut the required staffing ratio from 1 to 0.9 per three residents for facilities running multiple monitoring technologies", "stage": "mandate", "occurred_on": "2024-04-01", "verified_on": "2026-09-13", "source_name": "厚生労働省老健局——《令和6年度介護報酬改定における改定事項について》(全文 PDF)", "source_url": "https://www.mhlw.go.jp/content/12300000/001230329.pdf", "source_tier": "primary", "scope": "The ministry's own reform document for the 2024 fee revision, read directly. This is the rarest kind of record in this base: a government putting an exact number on how much technology substitutes for a person. For 特定施設 the required combined total of nursing and care staff moves from one per three residents to 0.9 per three — a ten per cent reduction in the staffing floor. It is conditional and the conditions are the substance. The facility must run multiple technologies, not one; must divide work between roles; must convene a committee including the multidisciplinary staff who actually deliver care; must trial the arrangement for at least three months while still meeting the normal staffing standard; and must submit data to the designating authority showing that care quality was maintained and staff burden reduced, with the approved staffing capped at what the trial established. A parallel change relaxes the night-shift add-on for dementia group homes from one extra full-time-equivalent to 0.9 where monitoring devices cover ten per cent of residents. Two limits. The safeguards list still requires individual night visits for residents who need them, so the rule does not treat a device as a substitute for going into the room. And this is a permission and a payment, not a measurement: nothing here counts a facility that took it up, a post reduced, or a worker who left. A separate new add-on pays facilities for adopting the technology at all, which is why this sits under policy rather than deployment.", "url": "https://flyvolo.ai/en/changes/ev-20240401-care-worker-2" }, { "id": "ev-20240401-construction-worker-2", "occupation_slug": "construction-worker", "task_ids": "setting-out; building-in-place", "title": "Japan's transport ministry stated a target of cutting the people needed on construction sites by at least 30% by fiscal 2040", "stage": "forecast", "occurred_on": "2024-04-01", "verified_on": "2026-09-13", "source_name": "国土交通省——《i-Construction 2.0 ～建設現場のオートメーション化～》(令和6年4月)", "source_url": "https://www.mlit.go.jp/tec/constplan/content/001738240.pdf", "source_tier": "primary", "scope": "A ministry strategy document, not a measurement, and its own framing matters: the target is stated as automation of the construction site, with at least a 30% reduction in people needed by fiscal 2040, equated to 1.5 times the productivity, alongside goals of reducing site fatalities and moving outdoor work to remote and off-site working. The context it gives is demographic rather than technological — Japan's working-age population is projected to fall by 20% by 2040, and construction already has a higher share of workers over 55 than the all-industry average. The part worth keeping is the admission the document makes against itself: it states that technologies for one person operating several machines at once, and for automating work on land and at sea, are not yet in general use, and that with the current measures productivity improvement has hit a ceiling. A target document saying the programme has plateaued is a different kind of evidence from one that does not. Nothing here counts a job, an hour or a site. It concerns Japan, and it is a goal a ministry set, not a change anyone has measured.", "url": "https://flyvolo.ai/en/changes/ev-20240401-construction-worker-2" }, { "id": "ev-20240401-social-worker-3", "occupation_slug": "social-worker", "task_ids": "risk-and-crisis", "title": "Allegheny County's own account of its child-welfare screening tool: supervisors override its default recommendation regularly, and the score is hidden once a case is opened", "stage": "deployment", "occurred_on": "2024-04-01", "verified_on": "2026-09-24", "source_name": "Allegheny County Department of Human Services — Summarizing Recent Research on Predictive Risk Models in Child Welfare (April 2024)", "source_url": "https://analytics.alleghenycounty.us/wp-content/uploads/2024/05/24-ACDHS-04-Predictive-Risk-Algorithms.pdf", "source_tier": "primary", "scope": "Allegheny County's own Department of Human Services published this review on its own analytics site, synthesising outside peer-reviewed and working-paper research about the Allegheny Family Screening Tool (AFST), which the county has used since 2016 to help call screening caseworkers decide whether a report of suspected child abuse or neglect is investigated. The document is dated April 2024 and was posted 31 May 2024 — a first-party account of how the tool is actually used, not a vendor's description of what it can do. Three statements in it carry the judgement for this task. First, on the decision itself: the county writes that AFST risk information is used \"alongside other information in the referral process,\" that intake supervisors \"have discretion to override the defaulted selections and do so regularly,\" and that supervisors historically overrode about a quarter of mandatory screen-in recommendations. Second, on what happens after screening: once a referral is screened in for investigation, \"neither the assigned investigator nor any caseworker ever sees the AFST score, so it does not affect ultimate assessments of maltreatment or decisions to open a case or remove a child from their home\" — the score never reaches the statutory removal decision at all. Third, citing De-Arteaga et al. (2020): during a period when a software defect caused some scores to display incorrectly, screeners' actual decisions tracked the true underlying risk more closely than the erroneous displayed scores, which the county reads as evidence that staff apply independent judgement rather than following the displayed number. What this does not establish: caseload, headcount, how long a screening decision takes, or how any of this generalises beyond one county's child-welfare hotline. Allegheny County built, funds and operates the AFST itself, so a claim about its own restraint in using it is not the same as an outside audit reaching the same conclusion — though the cited academic studies are independent of the county.", "url": "https://flyvolo.ai/en/changes/ev-20240401-social-worker-3" }, { "id": "ev-20240528-copywriter-1", "occupation_slug": "copywriter", "task_ids": "volume-copy", "title": "Klarna said generative AI now handles 80% of its copywriting through an internal 'Copy Assistant', and that AI accounted for about $10 million of annualised marketing savings in Q1 2024", "stage": "deployment", "occurred_on": "2024-05-28", "verified_on": "2026-09-10", "source_name": "Klarna — press release", "source_url": "https://www.klarna.com/international/press/ai-helps-klarna-cut-marketing-agency-spend-by-25-and-run-more-campaigns/", "source_tier": "primary", "scope": "One global consumer fintech's in-house marketing; company-reported figures with no independent audit. Applies to volume marketing copy, not to positioning work or regulated claims.", "url": "https://flyvolo.ai/en/changes/ev-20240528-copywriter-1" }, { "id": "ev-20240528-graphic-designer-1", "occupation_slug": "graphic-designer", "task_ids": "asset-production", "title": "Klarna said it generated 1,000+ marketing images with generative AI in Q1 2024, cutting the image development cycle from six weeks to seven days and image production costs by $6 million a year", "stage": "deployment", "occurred_on": "2024-05-28", "verified_on": "2026-09-10", "source_name": "Klarna — press release", "source_url": "https://www.klarna.com/international/press/ai-helps-klarna-cut-marketing-agency-spend-by-25-and-run-more-campaigns/", "source_tier": "primary", "scope": "In-house marketing imagery at one fintech (tools named: Midjourney, DALL-E, Firefly, Topaz Gigapixel, Photoroom). Campaign assets, not brand identity or bespoke illustration; company-reported.", "url": "https://flyvolo.ai/en/changes/ev-20240528-graphic-designer-1" }, { "id": "ev-20240528-marketing-specialist-1", "occupation_slug": "marketing-specialist", "task_ids": "content-and-creative-production", "title": "Klarna cut sales and marketing spend 11% in Q1 2024 while running more campaigns, attributing 37% of the savings (about $10 million annualised) to generative AI for images, copy and translation", "stage": "deployment", "occurred_on": "2024-05-28", "verified_on": "2026-09-10", "source_name": "Klarna — press release", "source_url": "https://www.klarna.com/international/press/ai-helps-klarna-cut-marketing-agency-spend-by-25-and-run-more-campaigns/", "source_tier": "primary", "scope": "One fintech's in-house marketing team; external agency spend fell 25%. Company-reported; the release gives no headcount figures.", "url": "https://flyvolo.ai/en/changes/ev-20240528-marketing-specialist-1" }, { "id": "ev-20240531-bus-driver-3", "occupation_slug": "bus-driver", "task_ids": "the-drive; remote-supervision; when-the-plan-breaks", "title": "Japan's Hokkaido Transport Bureau approved a vehicle on a Kamishihoro bus route as a Level 4 autonomous vehicle that needs no driver, for about 630 metres at up to about 12 km/h", "stage": "pilot", "occurred_on": "2024-05-31", "verified_on": "2026-09-23", "source_name": "MLIT Hokkaido District Transport Bureau — press release", "source_url": "https://wwwtb.mlit.go.jp/hokkaido/press/20240531_00002.html", "source_tier": "primary", "scope": "The transport bureau's own release and attached document, headed as a step toward solving the bus driver shortage. Under the Road Transport Vehicles Act it confirmed the automated driving system met the safety standard and granted operating-environment conditions: part of a town bus route of about 630 m, one way, operated by BOLDLY with a NAVYA ARMA vehicle, at up to about 12 km/h. The attached explainer states the Level 4 standard — when automated operation becomes difficult the vehicle must stop safely rather than hand over to a driver — and illustrates one remote monitor running three driverless vehicles; that ratio is a general illustration in the document, not a figure for this route. It is recorded as a pilot rather than a deployment because of its scale, and it establishes a regulatory approval, not ridership or any change in driver numbers.", "url": "https://flyvolo.ai/en/changes/ev-20240531-bus-driver-3" }, { "id": "ev-20240701-auditor-2", "occupation_slug": "auditor", "task_ids": "the-file-and-the-opinion; owning-machine-drafted-work", "title": "PCAOB staff reported large audit firms' generative-AI use sitting in administrative and research work, with some firms barring it from audit procedures", "stage": "deployment", "occurred_on": "2024-07-01", "verified_on": "2026-09-12", "source_name": "Public Company Accounting Oversight Board — Spotlight: Staff Update on Outreach Activities Related to the Integration of Generative Artificial Intelligence in Audits and Financial Reporting", "source_url": "https://pcaobus.org/documents/generative-ai-spotlight.pdf", "source_tier": "primary", "scope": "A snapshot from the US audit regulator's own outreach, and the firms are the large end of the market: the US global network firms plus non-affiliated firms auditing more than 100 issuers, which the report says collectively audit the majority of issuer market capitalisation. What was deployed is narrow — staff drafting administrative documents and initial memos, and firm-built tools for researching internal accounting and auditing guidance. The report states current integration 'appears to be focused primarily on administrative and research activities', that some firms do not allow generative AI in audit or attest procedures at all, and that firms expect it to augment rather than replace. It is a regulator relaying what firms told it, not an inspection finding, and it dates from July 2024 in an area the report itself calls rapidly evolving; smaller firms were described as further behind.", "url": "https://flyvolo.ai/en/changes/ev-20240701-auditor-2" }, { "id": "ev-20240702-product-designer-1", "occupation_slug": "product-designer", "task_ids": "screen-production", "title": "Figma disabled its Make Designs prompt-to-UI feature a week after launch when 'weather app' prompts produced screens closely resembling Apple's; the CEO cited insufficient QA; relaunched Sept 2024", "stage": "constraint", "occurred_on": "2024-07-02", "verified_on": "2026-09-10", "source_name": "Figma — blog retrospective", "source_url": "https://www.figma.com/blog/inside-figma-a-retrospective-on-make-designs/", "source_tier": "primary", "scope": "One design-tool vendor's generative feature (GPT-4o and Amazon Titan on hand-built component design systems), limited beta. A vendor rollback, not a customer deployment; the feature returned within three months as First Draft.", "url": "https://flyvolo.ai/en/changes/ev-20240702-product-designer-1" }, { "id": "ev-20240711-actuary-1", "occupation_slug": "actuary", "task_ids": "the-price; fairness-testing", "title": "New York's financial regulator told insurers to test AI systems and external data used in underwriting and pricing for unfair discrimination, before production and regularly afterwards", "stage": "constraint", "occurred_on": "2024-07-11", "verified_on": "2026-09-23", "source_name": "New York State Department of Financial Services — Insurance Circular Letter No. 7 (2024)", "source_url": "https://www.dfs.ny.gov/industry-guidance/circular-letters/cl2024-07", "source_tier": "primary", "scope": "The regulator's own circular letter of 11 July 2024, addressed to insurers authorised in New York, fraternal societies, HMOs and the State Insurance Fund. Its verbs are expectations (should), not a statute, and it covers underwriting and pricing only. It expects unfair or unlawful discrimination testing before an AI system goes into production and on a regular cadence thereafter, lists quantitative methods including the adverse impact ratio, denial odds ratios and marginal effects, expects insurers to show that external data are supported by generally accepted actuarial standards of practice, requires comprehensive documentation, places oversight on the board and senior management, and says insurers retain responsibility for third-party vendor tools. It does not say who inside an insurer does the testing, and it measures nothing about actuaries' work or numbers.", "url": "https://flyvolo.ai/en/changes/ev-20240711-actuary-1" }, { "id": "ev-20240801-backend-developer-1", "occupation_slug": "backend-developer", "task_ids": "endpoints-and-plumbing; cost-latency-capacity", "title": "Amazon says it moved tens of thousands of its own production Java applications from Java 8 or 11 to Java 17 with an agent, and estimates the manual equivalent at over 4,500 years of development work", "stage": "deployment", "occurred_on": "2024-08-01", "verified_on": "2026-09-12", "source_name": "AWS DevOps & Developer Productivity Blog (Amazon's own)", "source_url": "https://aws.amazon.com/blogs/devops/amazon-q-developer-just-reached-a-260-million-dollar-milestone/", "source_tier": "primary", "scope": "Amazon's own internal estimate about Amazon's own engineering, published by the company that sells the tool - the party doing the work, the party measuring it and the party selling it are the same, which is stated here because it cannot be checked from outside. The company also publishes its method: time saved was estimated from the number of Java dependencies migrated, assuming a day or more of developer time per dependency manually. Note what the migration is: a language-version upgrade has a compiler and an existing test suite as its oracle, so success is machine-checkable at every step - the most favourable possible shape for this kind of automation and not a general result about backend work. And note what is claimed and what is not: 4,500 years is work not done, over a thousand developers, with no statement anywhere that any of them left or that a role was removed.", "url": "https://flyvolo.ai/en/changes/ev-20240801-backend-developer-1" }, { "id": "ev-20240801-hr-recruiter-1", "occupation_slug": "hr-recruiter", "task_ids": "sourcing-and-screening", "title": "The EU AI Act (in force 1 August 2024) lists AI used to recruit or select people — placing targeted job ads, filtering applications, evaluating candidates — as high-risk under Annex III, point 4", "stage": "constraint", "occurred_on": "2024-08-01", "verified_on": "2026-09-11", "source_name": "Regulation (EU) 2024/1689 (Artificial Intelligence Act), Annex III point 4 — verbatim mirror of the Official Journal text", "source_url": "https://artificialintelligenceact.eu/annex/3/", "source_tier": "primary", "scope": "EU market. Annex III obligations (risk management, human oversight, transparency) were scheduled to apply from 2 August 2026 and are subject to amendment. Source mirrors the regulation text; the Official Journal text is Regulation (EU) 2024/1689.", "url": "https://flyvolo.ai/en/changes/ev-20240801-hr-recruiter-1" }, { "id": "ev-20240804-video-editor-3", "occupation_slug": "video-editor", "task_ids": "assembly-and-rough-cut", "title": "IATSE's 2024 Basic Agreement added Article XLIX, keeping work done by prompting or overseeing an AI system inside covered union work", "stage": "constraint", "occurred_on": "2024-08-04", "verified_on": "2026-09-11", "source_name": "IATSE — 2024 Basic Agreement MOA, Article XLIX (fully executed; copy hosted by IATSE Local 728)", "source_url": "https://www.iatse728.org/sites/default/files/files/MOA%20-%202024%20IATSE%20Basic%20Agreement%20-%20Fully%20Executed.PDF", "source_tier": "primary", "scope": "US union film and television production under the IATSE Basic Agreement, applying from 4 August 2024; the Videotape Agreement incorporates the same Article by reference. The Article is craft-agnostic, so it reaches picture editing as bargaining-unit work. It also bars producers from requiring employee-furnished prompts in a way that displaces a covered employee. It does not cover non-union production, advertising or social video, and it does not stop a producer requiring AI use — it keeps that use inside the contract.", "url": "https://flyvolo.ai/en/changes/ev-20240804-video-editor-3" }, { "id": "ev-20240919-data-analyst-1", "occupation_slug": "data-analyst", "task_ids": "query-writing", "title": "Uber put QueryGPT, a natural-language-to-SQL tool, into production for operations and support teams, reporting query authoring time down from about 10 to about 3 minutes (~300 daily users)", "stage": "deployment", "occurred_on": "2024-09-19", "verified_on": "2026-09-10", "source_name": "Uber — engineering blog", "source_url": "https://www.uber.com/en-US/blog/query-gpt/", "source_tier": "primary", "scope": "One company, limited release; Uber's platform runs about 1.2 million interactive queries a month, so this is a small share. The post itself flags hallucinated tables and columns and prompt quality as open problems.", "url": "https://flyvolo.ai/en/changes/ev-20240919-data-analyst-1" }, { "id": "ev-20241001-farmer-2", "occupation_slug": "farmer", "task_ids": "field-work", "title": "Japan's Smart Agriculture Act took effect, funding farmers who adopt smart-farming technology through two certification schemes", "stage": "mandate", "occurred_on": "2024-10-01", "verified_on": "2026-09-23", "source_name": "Ministry of Agriculture, Forestry and Fisheries of Japan — スマート農業技術活用促進法について", "source_url": "https://www.maff.go.jp/j/kanbo/smart/houritsu.html", "source_tier": "primary", "scope": "The ministry's own page on the Act, read in full. The Act on Promoting the Use of Smart Agricultural Technology to Improve Agricultural Productivity was enacted on 14 June 2024, promulgated on 21 June and took effect on 1 October 2024. In response to changes including the decline in the number of farmers, it creates two certification schemes — a production-method innovation plan for farmers who adopt smart-farming technology together with new ways of producing, and a development and supply plan for businesses that develop and spread such technology — and certified farmers and businesses can receive financial and other support; the page adds priority in subsidy programmes and a tax measure. It is a law that funds adoption; it does not show how many farmers were certified or what adoption followed.", "url": "https://flyvolo.ai/en/changes/ev-20241001-farmer-2" }, { "id": "ev-20241001-port-worker-2", "occupation_slug": "port-worker", "task_ids": "what-the-contract-says; crane-work", "title": "The master contract covering US east and gulf coast ports bars fully automated terminals and remotely operated ship-to-shore cranes for its six-year term, and sets one operator per machine", "stage": "constraint", "occurred_on": "2024-10-01", "verified_on": "2026-09-20", "source_name": "USMX–ILA Master Contract, effective 1 October 2024 for a six-year term — Article XI, New Technology Implementation and Workforce Protection", "source_url": "https://usmx.com/wp-content/uploads/2026/08/2024-2030_USMX-ILA_Master_Contract.pdf", "source_tier": "primary", "scope": "United States east and gulf coast ports only, and only for the term of this agreement, which runs from 1 October 2024 to 30 September 2030. It is an executed collective agreement published by a signatory, so both parties have an interest and they point in opposite directions, which is what makes the text stronger than either side's account of it. Article XI says there shall be no fully-automated terminals developed and no fully-automated container handling equipment used during the term, defining fully-automated as devoid of human interaction, and separately that no remotely operated ship-to-shore cranes shall be used at any port during the term. Rail-mounted and rubber-tyred gantry cranes may be operated remotely, but the clause requires that the equipment provide the represented operator with control of the vertical and horizontal movement of cargoes for vessel loading and unloading and truck receiving and delivery, with a minimum manning of one operator per machine while gate and vessel operations run simultaneously. The workforce-protection procedure requires the parties to identify the new work created by the technology by craft before implementation, which is unusually close to the unit this site uses. Enforcement is real rather than hortatory: a terminal found in violation pays ten thousand dollars per day to the local container royalty fund, and where it refuses to acknowledge a violation the union may withhold labour. What it does not establish: nothing about whether the technology works, and nothing about any other country. It is a bargained outcome, and reading it as evidence that automation failed inverts it — it is evidence that where dockworkers had the leverage to bargain, the manning was written down rather than engineered away. It also expires.", "url": "https://flyvolo.ai/en/changes/ev-20241001-port-worker-2" }, { "id": "ev-20241009-machine-learning-engineer-1", "occupation_slug": "machine-learning-engineer", "task_ids": "building-a-model-for-the-problem; deciding-good-enough", "title": "OpenAI built a benchmark of 75 Kaggle competitions to test agents at ML engineering; the best setup reached bronze-medal level in 16.9% of them", "stage": "capability", "occurred_on": "2024-10-09", "verified_on": "2026-09-12", "source_name": "MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering (OpenAI, arXiv:2410.07095)", "source_url": "https://arxiv.org/abs/2410.07095", "source_tier": "primary", "scope": "Seventy-five ML-engineering Kaggle competitions - training models, preparing datasets, running experiments - graded against the human baselines on each competition's own public leaderboard, so the comparison is with the people who actually entered. The best-performing setup at the time of writing (o1-preview with the AIDE scaffold) reached at least bronze in 16.9% of competitions; the authors also report that resource scaling and pre-training contamination both change the number, and the benchmark is open-sourced so anyone can rerun it on a newer model. Two limits are structural rather than incidental: a Kaggle competition arrives with the problem already framed, the metric already chosen and the data already collected, which is the part of this job the benchmark cannot test at all; and OpenAI built the benchmark and sells models measured on it.", "url": "https://flyvolo.ai/en/changes/ev-20241009-machine-learning-engineer-1" }, { "id": "ev-20241029-junior-software-developer-1", "occupation_slug": "junior-software-developer", "task_ids": "boilerplate; review-generated", "title": "Alphabet's CEO said on the Q3 2024 earnings call that more than a quarter of all new code at Google is generated by AI, then reviewed and accepted by engineers", "stage": "deployment", "occurred_on": "2024-10-29", "verified_on": "2026-09-10", "source_name": "Google — CEO remarks, Alphabet Q3 2024 earnings call", "source_url": "https://blog.google/company-news/inside-google/message-ceo/alphabet-earnings-q3-2024/", "source_tier": "primary", "scope": "One large technology company, company-reported, with no definition of how 'generated' is measured (autocomplete versus whole functions). Says nothing about hiring; the review step stayed with engineers.", "url": "https://flyvolo.ai/en/changes/ev-20241029-junior-software-developer-1" }, { "id": "ev-20241112-data-engineer-1", "occupation_slug": "data-engineer", "task_ids": "building-the-pipeline; what-the-number-means", "title": "On 632 enterprise data workflows drawn from real warehouses, a code agent solved 21.3% — against 91.2% on the academic version of the same task", "stage": "constraint", "occurred_on": "2024-11-12", "verified_on": "2026-09-12", "source_name": "Spider 2.0: Evaluating Language Models on Real-World Enterprise Text-to-SQL Workflows (arXiv:2411.07763)", "source_url": "https://arxiv.org/abs/2411.07763", "source_tier": "primary", "scope": "632 problems built from real data applications on BigQuery and Snowflake, with databases often exceeding 1,000 columns. What makes them hard is stated by the authors and is exactly this occupation's daily environment: the answer requires searching database metadata and dialect documentation, reading the project's own codebase, holding very long context, and emitting several queries in different dialects that frequently run past 100 lines. The 21.3% is one agent framework on o1-preview at the end of 2024 and will be out of date; the durable finding is the gap to 91.2% on Spider 1.0 and 73.0% on BIRD, which is a measurement of how much of the difficulty lives in the warehouse rather than in the SQL. It says nothing about employment, and nothing about pipelines that are built rather than queried - it tests producing the query, not operating the system that keeps producing it.", "url": "https://flyvolo.ai/en/changes/ev-20241112-data-engineer-1" }, { "id": "ev-20241112-video-editor-1", "occupation_slug": "video-editor", "task_ids": "generative-and-hybrid-production", "title": "Coca-Cola's 2024 holiday campaign reimagined its 1995 'Holidays Are Coming' spot with films made by three AI production studios using generative video tools", "stage": "deployment", "occurred_on": "2024-11-12", "verified_on": "2026-09-11", "source_name": "The Coca-Cola Company — media centre, 2024 holiday campaign", "source_url": "https://www.coca-colacompany.com/media-center/groundbreaking-digital-experience-and-films-fuse-holiday-heritage-with-cutting-edge-tech", "source_tier": "primary", "scope": "One global brand's seasonal campaign. The company names the three studios — Secret Level (Los Angeles), The Wild Card (Kuala Lumpur) and Silverside AI Tech Lab (San Francisco) — and technology partners including Microsoft, Leonardo.ai and OpenAI. The AI spots ran alongside conventionally produced advertising and drew public criticism from creative professionals, which the company release does not mention. It says nothing about how many editors worked on the films, what they were paid, or whether the brand spent less; a brand reimagining one archive spot is not evidence about commercial video production generally.", "url": "https://flyvolo.ai/en/changes/ev-20241112-video-editor-1" }, { "id": "ev-20241121-airline-pilot-3", "occupation_slug": "airline-pilot", "task_ids": "the-new-category", "title": "The FAA's final rule creating Part 194 requires a type rating for any pilot in command of a powered-lift (eVTOL) aircraft", "stage": "constraint", "occurred_on": "2024-11-21", "verified_on": "2026-09-25", "source_name": "U.S. Federal Register — FAA final rule, Integration of Powered-Lift: Pilot Certification and Operations (21 Nov 2024)", "source_url": "https://www.govinfo.gov/content/pkg/FR-2024-11-21/html/2024-24886.htm", "source_tier": "primary", "scope": "U.S. Federal Aviation Administration final rule (Docket No. FAA-2023-1275; RIN 2120-AL72), published in the Federal Register on 21 November 2024 and effective 21 January 2025 (one amendatory instruction effective 21 July 2025). It adds 14 CFR part 194, Special Federal Aviation Regulation No. 120, in force for ten years. The rule's own purpose section states that existing part 61 training and certification rules for pilots and flight instructors \"do not adequately address the unique challenges of introducing a new category of aircraft to civil operations,\" and its executive summary notes that \"currently, there are no type-certificated powered-lift in civil operations\" (powered-lift is the FAA's own regulatory term for what is commonly called an eVTOL aircraft: heavier-than-air, capable of vertical takeoff, vertical landing and low-speed flight). The rule requires every pilot in command of a powered-lift to hold a type rating for that specific aircraft, and, because in the FAA's words \"once the first powered-lift achieve type certification, there will be an insufficient number of qualified flight instructors,\" it creates an alternate training pathway through the manufacturer's own test flights for the initial cadre of instructors and pilots. This is United States federal civil aviation rulemaking only: it says nothing about any aircraft actually operating commercially yet (none was type-certificated as of this rule), nothing about the operating rules in part 135 beyond certification and training, and nothing about any other country's regulation of the same aircraft category.", "url": "https://flyvolo.ai/en/changes/ev-20241121-airline-pilot-3" }, { "id": "ev-20241201-compliance-officer-1", "occupation_slug": "compliance-officer", "task_ids": "the-alert-queue; the-evidence-file", "title": "US Treasury reported that firms told it AI is widely used for AML and sanctions compliance, and that generative AI has been deployed to automate report creation and filing", "stage": "deployment", "occurred_on": "2024-12-01", "verified_on": "2026-09-13", "source_name": "U.S. Department of the Treasury, Artificial Intelligence in Financial Services (report on the June 2024 RFI)", "source_url": "https://home.treasury.gov/system/files/136/Artificial-Intelligence-in-Financial-Services.pdf", "source_tier": "primary", "scope": "US financial services, published December 2024, synthesising comment letters submitted to a Treasury request for information issued in June 2024 and closed that August. Read what kind of evidence this is: it is a government department reporting what respondents told it, not a survey with a sampling frame and not a count of installations — anyone can write a comment letter and the firms that do are the ones with something to say. What makes it worth recording anyway is that it is the authority of record publishing under its own name, that the uses named are specific rather than aspirational (anomaly detection, flagging suspicious activity, identity verification under Bank Secrecy Act obligations, and collating and summarising data inside an investigation platform), and that report creation and filing — the evidence half of this job — is named as already deployed rather than planned.", "url": "https://flyvolo.ai/en/changes/ev-20241201-compliance-officer-1" }, { "id": "ev-20241231-administrative-assistant-1", "occupation_slug": "administrative-assistant", "task_ids": "scheduling-and-correspondence", "title": "A UK cross-government trial gave 20,000 civil servants in 12 organisations Microsoft 365 Copilot for three months; users self-reported saving 26 minutes a day on average, mostly on drafting", "stage": "pilot", "occurred_on": "2024-12-31", "verified_on": "2026-09-10", "source_name": "UK Government Digital Service — M365 Copilot Experiment findings report", "source_url": "https://www.gov.uk/government/publications/microsoft-365-copilot-experiment-cross-government-findings-report", "source_tier": "primary", "scope": "UK central government, 30 Sep–31 Dec 2024, self-reported time savings with no control group; scheduling meetings saved about 9 minutes, drafting documents 24. The report notes the tool struggled with complex data and that human oversight was needed throughout.", "url": "https://flyvolo.ai/en/changes/ev-20241231-administrative-assistant-1" }, { "id": "ev-20250130-bank-teller-1", "occupation_slug": "bank-teller", "task_ids": "routine-transactions", "title": "Lloyds Banking Group announced 136 UK branch closures for May 2025–March 2026, citing 10 million fewer branch visits in 2024 and a 48% five-year fall in branch transactions", "stage": "labor_impact", "occurred_on": "2025-01-30", "verified_on": "2026-09-10", "source_name": "PYMNTS", "source_url": "https://www.pymnts.com/news/banking/2025/lloyds-shuttering-136-branches-amid-digital-banking-shift", "source_tier": "secondary", "scope": "UK retail banking (Lloyds, Halifax, Bank of Scotland). Affected staff were offered other roles: a change in where and what branch staff do, not announced redundancies.", "url": "https://flyvolo.ai/en/changes/ev-20250130-bank-teller-1" }, { "id": "ev-20250224-lawyer-3", "occupation_slug": "lawyer", "task_ids": "drafting-and-negotiation", "title": "A Wyoming federal court fined three attorneys and revoked one's pro hac vice admission after a motion they filed cited nine cases, eight of which did not exist", "stage": "constraint", "occurred_on": "2025-02-24", "verified_on": "2026-09-11", "source_name": "US District Court, D. Wyo. — Wadsworth v. Walmart, 2:23-cv-118-KHR, ECF No. 181 (order of Judge Kelly H. Rankin)", "source_url": "https://storage.courtlistener.com/recap/gov.uscourts.wyd.64014/gov.uscourts.wyd.64014.181.0_1.pdf", "source_tier": "primary", "scope": "One US federal civil case. The motions were filed 22 January 2025 and withdrawn the day after the show-cause order; the attorneys admitted the cases were hallucinated by an AI platform. The drafter was fined USD 3,000 and lost his pro hac vice admission; the supervising and local attorneys were fined USD 1,000 each. In setting the amount the court weighed that attorneys have been on notice of this failure mode for some time. The sanction falls on the signing and verification duty, not on using AI as such.", "url": "https://flyvolo.ai/en/changes/ev-20250224-lawyer-3" }, { "id": "ev-20250314-ecommerce-operator-2", "occupation_slug": "ecommerce-operator", "task_ids": "the-listing", "title": "Four Chinese ministries issued labelling rules placing the duty to declare AI-generated content on the person publishing it, not only on the tool provider", "stage": "constraint", "occurred_on": "2025-03-14", "verified_on": "2026-09-13", "source_name": "国家互联网信息办公室等四部门——《人工智能生成合成内容标识办法》(全文)", "source_url": "https://www.cac.gov.cn/2025-03/14/c_1743654684782215.htm", "source_tier": "primary", "scope": "The full fourteen articles read on the Cyberspace Administration's own site, issued jointly with the Ministry of Industry and Information Technology, the Ministry of Public Security and the National Radio and Television Administration on 14 March 2025 and in force from 1 September 2025; a mandatory national standard on labelling methods took effect the same day. Most of the measure binds service providers — Article 4 requires a visible label on generated text, audio, images, video and virtual scenes, and Article 5 requires an implicit label carrying provider and content identifiers in the file metadata. Two articles reach past the vendor to the person publishing. Article 10 requires a user who posts generated content through a distribution service to declare it and use the labelling function provided, and forbids anyone from maliciously deleting, altering, forging or concealing a label, or supplying tools to do so. Article 9 allows a user to obtain content without a visible label only after the provider has made the user's own labelling duty explicit in the agreement, and the provider must then keep logs of who received it for at least six months. Article 6 makes the distribution platform verify metadata and add a visible label itself, including where it merely detects traces. This is Chinese law and applies to services provided within China. It does not say how many listings use generated copy, and it does not establish that anyone was penalised — the measure leaves enforcement to the relevant authorities under existing law.", "url": "https://flyvolo.ai/en/changes/ev-20250314-ecommerce-operator-2" }, { "id": "ev-20250314-frontend-developer-1", "occupation_slug": "frontend-developer", "task_ids": "owning-a-design-system", "title": "Airbnb reports migrating nearly 3,500 React component test files from Enzyme to React Testing Library in six weeks, against a hand estimate of 1.5 years", "stage": "deployment", "occurred_on": "2025-03-14", "verified_on": "2026-09-12", "source_name": "The Airbnb Tech Blog (Airbnb's own engineering blog)", "source_url": "https://medium.com/airbnb-engineering/accelerating-large-scale-test-migration-with-llms-9565c208023b", "source_tier": "primary", "scope": "One company's own codebase, published by that company, with the hand-migration estimate (1.5 years of engineering time) supplied by the same company and not checkable from outside. What makes the result reproducible is the mechanism, which the post describes in full: the migration is a state machine whose every step has an automatic referee - jest, eslint and tsc say pass or fail - so the model could retry without a person watching, and the first bulk run cleared 75% of files in four hours. The remaining tail is the more useful number: after four days of tuning, 97%; the last 3% had each been retried between 50 and 100 times and were finished by hand. The post is also explicit that the main driver was selecting the right related files to put in the prompt - which grew to 40,000-100,000 tokens pulling in up to 50 files - rather than prompt wording. It reports no change to headcount, hiring or roles, and does not claim one.", "url": "https://flyvolo.ai/en/changes/ev-20250314-frontend-developer-1" }, { "id": "ev-20250318-accountant-1", "occupation_slug": "accountant", "task_ids": "compliance-filing; data-entry", "title": "EY announced an initial deployment of 150 AI agents to support 80,000 of its tax professionals in data collection, document review and income and indirect tax compliance, built with NVIDIA", "stage": "deployment", "occurred_on": "2025-03-18", "verified_on": "2026-09-10", "source_name": "EY — press release", "source_url": "https://www.ey.com/en_us/newsroom/2025/03/ey-launching-ey-ai-agentic-platform-created-with-nvidia-ai-to-drive-multi-sector-transformation-starting-with-tax-risk-and-finance-domains", "source_tier": "primary", "scope": "One Big Four firm's own tax practice, global; a launch announcement whose scale figures (3 million tax deliverables, 30 million processes 'over the coming year') are forward-looking targets, not measured results. No headcount statement.", "url": "https://flyvolo.ai/en/changes/ev-20250318-accountant-1" }, { "id": "ev-20250326-air-traffic-controller-3", "occupation_slug": "air-traffic-controller", "task_ids": "the-sequence", "title": "NATS deployed its Intelligent Approach sequencing tool at Gatwick Airport, giving controllers radar markers to sequence arrivals from live wind data instead of fixed spacing distances", "stage": "deployment", "occurred_on": "2025-03-26", "verified_on": "2026-09-23", "source_name": "NATS (UK air navigation service provider) — press release", "source_url": "https://www.nats.aero/news/nats-introduces-world-first-aircraft-separation-system-at-gatwick-cutting-carbon-emissions-and-boosting-on-time-performance", "source_tier": "primary", "scope": "NATS (National Air Traffic Services), the UK's designated air navigation service provider, announced on its own news site (26 March 2025) that it had deployed Intelligent Approach with Advanced Mixed Mode capability, a tool developed jointly with Leidos, at London Gatwick Airport, the first time it has been used at a single-runway mixed-mode airport where arrivals and departures share a runway. The tool replaces fixed-distance aircraft separation with dynamically calculated time-based intervals based on live wind conditions, and works by putting markers on a controller's radar screen that help them precisely place each arrival to build the most efficient sequence; NATS states controllers keep the decision. NATS reports the tool had already been deployed at Heathrow, where it cut headwind delays by 62 percent, and at Toronto Pearson and Amsterdam Schiphol. For Gatwick, NATS predicts the change will cut CO2 emissions by 11,000 to 19,000 tonnes a year and reduce holding and delays; those Gatwick figures are a prediction stated alongside the deployment, not a separately measured outcome. This describes one tool at four named airports, developed and reported by the air navigation provider that operates it, and does not establish adoption at other providers or say what share of a controller's shift the tool now covers.", "url": "https://flyvolo.ai/en/changes/ev-20250326-air-traffic-controller-3" }, { "id": "ev-20250331-farmer-1", "occupation_slug": "farmer", "task_ids": "the-spraying", "title": "Japan's agriculture ministry estimated drones sprayed about 1.2 million hectares of cumulative farmland area in fiscal 2024", "stage": "deployment", "occurred_on": "2025-03-31", "verified_on": "2026-09-23", "source_name": "Ministry of Agriculture, Forestry and Fisheries of Japan — 令和7年度 農業分野におけるドローンの活用状況 (Crop Production Bureau, March 2026)", "source_url": "https://www.maff.go.jp/j/kanbo/smart/attach/pdf/drone-205.pdf", "source_tier": "primary", "scope": "The ministry's own report on drone use in agriculture, dated March 2026, read in full. It states that the cumulative area sprayed with pesticide, fertiliser and similar materials by drone passed one million hectares in fiscal 2023 and continued to expand, at about 1,196 thousand hectares in fiscal 2024 (an estimate); that sales of spraying drones have run at roughly 3,000 to 4,000 a year, rising slightly to 3,359 in fiscal 2024; and that 1,461 pesticides suitable for drone spraying were registered by the end of fiscal 2024. Its own method note says the area was compiled from operators' reports in 2016–2018, estimated from prefectural data in 2019 and from flight-time reports in 2020, not compiled in 2021, and estimated from registered aircraft and average area per aircraft from 2022 onward; the series therefore mixes methods and is not a reliable growth rate. Cumulative area counts a field once per spraying. It establishes the scale of drone spraying in one country; it does not count farmers, or say how many use drones themselves versus hiring a service.", "url": "https://flyvolo.ai/en/changes/ev-20250331-farmer-1" }, { "id": "ev-20250401-receptionist-2", "occupation_slug": "receptionist", "task_ids": "signing-people-in; whatever-walks-in", "title": "Japan's health ministry set out when a lodging front desk may go unstaffed, and the condition is that a member of staff can reach the guest in roughly ten minutes", "stage": "constraint", "occurred_on": "2025-04-01", "verified_on": "2026-09-13", "source_name": "厚生労働省健康・生活衛生局長通知 健生発0311第1号——《旅館業における衛生等管理要領の一部改正について》", "source_url": "https://www.mhlw.go.jp/web/t_doc?dataId=00tc9552&dataType=1&pageNo=1", "source_tier": "primary", "scope": "The ministry's own notice to prefectural governors and public health authorities, applying from 1 April 2025. Two layers have to be kept apart or the record overstates the law. The binding layer is the ordinance: article 4-3 of the Lodging Business Act enforcement regulation lets equipment stand in for a front desk only where it enables rapid response to an accident or other emergency, and enables the guest register to be kept accurately, room keys to be handed over properly, and the entry and exit of non-guests to be confirmed. The ordinance sets no time. The number comes from the ministry's management guideline, which describes a system in which, on request, staff can normally reach a guest in an urgent situation in about ten minutes, and which also expects identity and entry checks to be carried out through continuously clear images from cameras the operator installed themselves. Read the notice's own last line before quoting any of it: it states that it is technical advice under article 245-4(1) of the Local Autonomy Act, and it asks prefectures to apply it flexibly to local circumstances and amend their ordinances as needed. So the ten minutes is an expectation transmitted through local regulation, not a national rule with a penalty attached. It applies in Japan, and it counts nothing — no unstaffed desks, no response times, no posts.", "url": "https://flyvolo.ai/en/changes/ev-20250401-receptionist-2" }, { "id": "ev-20250401-sales-account-manager-3", "occupation_slug": "sales-account-manager", "task_ids": "research-and-qualification; operating-the-selling-system", "title": "Singapore's financial regulator and the sector's skills bodies forecast that Gen AI will augment most financial-sector roles over five years rather than displace them", "stage": "forecast", "occurred_on": "2025-04-01", "verified_on": "2026-09-12", "source_name": "Generative AI Jobs Transformation Map for Singapore's financial sector (MAS, IBF and Workforce Singapore)", "source_url": "https://www.ibf.org.sg/docs/default-source/jtm/genai-jtm.pdf", "source_tier": "primary", "scope": "Singapore's financial sector only, across retail, corporate, investment and private banking, wealth and asset management, and insurance. It is an expectation, not a measurement: the study examines likely adoption trends and the skills the workforce would need, and states plainly that it is difficult to predict with certainty how Gen AI will affect the workforce. Two of its specifics are what make it worth recording beside the US data. First, it expects about 80% of Singapore's potential value to come from four business functions — sales and marketing, customer operations, risk management, and engineering and technology — which is independent of, and largely agrees with, the US Census finding that adoption concentrates in sales and marketing and IT. Second, it classifies roles by whether they 'do more' or 'do more and do new' rather than by whether they survive. The analysis was contracted to a consultancy that sells AI transformation work, and the commissioning bodies have a stake in the sector remaining competitive; both belong beside the conclusion.", "url": "https://flyvolo.ai/en/changes/ev-20250401-sales-account-manager-3" }, { "id": "ev-20250403-ai-implementation-lead-1", "occupation_slug": "ai-implementation-lead", "task_ids": "choosing-what-to-try-first; setting-the-guardrails", "title": "US OMB Memorandum M-25-21 requires every executive department and agency, including independent regulators, to identify a Chief AI Officer and allocate resources and responsibilities to the role", "stage": "mandate", "occurred_on": "2025-04-03", "verified_on": "2026-09-11", "source_name": "US Office of Management and Budget, Memorandum M-25-21 (3 April 2025), signed by OMB Director Russell T. Vought", "source_url": "https://www.whitehouse.gov/wp-content/uploads/2025/02/M-25-21-Accelerating-Federal-Use-of-AI-through-Innovation-Governance-and-Public-Trust.pdf", "source_tier": "primary", "scope": "The United States federal executive branch only, not private employers. What it establishes is that this role is required to exist by rule rather than created at each organisation's discretion. Three things need saying. First, the memorandum rescinds and replaces 2024's M-24-10 and deliberately redefines the role: it describes the Chief AI Officer as a champion of adoption rather than a layer of oversight, so this record supports the adoption-driving side of the job, not the gatekeeping side. Second, it requires an agency to identify an officer — not a headcount, a budget or a team; an agency may name its existing chief information officer. Third, a rule requiring a post to exist and that post actually doing the work are two different layers, and this site records it at the policy layer, where one record changes no task judgement on its own.", "url": "https://flyvolo.ai/en/changes/ev-20250403-ai-implementation-lead-1" }, { "id": "ev-20250409-electrician-1", "occupation_slug": "electrician", "task_ids": "electrification-work", "title": "US contractor Rosendin reported field trials of a three-robot solar module installer: robots plus a two-person crew set 350–400 modules per 8-hour shift; electricians do the grid connections", "stage": "pilot", "occurred_on": "2025-04-09", "verified_on": "2026-09-10", "source_name": "pv magazine (reporting Rosendin's announcement)", "source_url": "https://www.pv-magazine.com/2025/04/09/robots-to-work-side-by-side-with-humans-to-demonstrate-solar-module-installation/", "source_tier": "secondary", "scope": "Utility-scale solar fields in Texas, module placement only (developed with ULC Technologies); demonstrated on site on 17 April 2025. Not building wiring or residential work; contractor-reported rates with no independent measurement.", "url": "https://flyvolo.ai/en/changes/ev-20250409-electrician-1" }, { "id": "ev-20250415-experienced-software-engineer-3", "occupation_slug": "experienced-software-engineer", "task_ids": "reviewing-what-the-machine-wrote; owning-the-agent-that-writes", "title": "US Immigration and Customs Enforcement reported deploying AI developer tools on 15 April 2025, and recorded that their output must be reviewed and approved before it enters the codebase", "stage": "deployment", "occurred_on": "2025-04-15", "verified_on": "2026-09-20", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory (entry DHS-2758)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory", "source_tier": "primary", "scope": "One agency component (ICE) inside one government, reported by that agency under a statutory duty rather than chosen for publication, which is what makes the boring entries in this inventory worth reading. Entry DHS-2758 in the individually-reported CSV. What it establishes is the deployment and the conditions attached to it, not how much the tools are used: the inventory carries no usage figure, no headcount and no productivity claim beyond the agency saying the tools \"increase developer productivity\". The two clauses that support the two linked tasks are separate sentences in the words of the agency itself: outputs \"must be reviewed and approved before being incorporated into the codebase through existing version control and deployment processes\" (the reviewing task), and \"the tools do not directly modify production systems; all changes must go through standard human review, testing, and approval workflows\" (the task of setting what an agent may do unattended). The named tool is a vendor purchase, which is worth knowing though the agency, not the vendor, is the party reporting here. The same inventory carries a counter-signal from the same department: entry DHS-373, \"Commercial Generative AI for Code Generation\", has a development stage of Retired. That row carries no date and no reason, so it establishes only that DHS listed such a use case and no longer runs it.", "url": "https://flyvolo.ai/en/changes/ev-20250415-experienced-software-engineer-3" }, { "id": "ev-20250422-delivery-rider-1", "occupation_slug": "delivery-rider", "task_ids": "the-ride", "title": "Meituan received China's first nationwide low-altitude logistics operating certificate from the CAAC for its delivery drones, after 450,000+ drone orders on 53 routes by end-2024", "stage": "deployment", "occurred_on": "2025-04-22", "verified_on": "2026-09-10", "source_name": "China Daily", "source_url": "https://www.chinadaily.com.cn/a/202504/22/WS6807752fa3104d9fd3820e14.html", "source_tier": "secondary", "scope": "China; routes concentrated in Shenzhen, Beijing, Shanghai, Guangzhou and Nanjing. Volume is a small fraction of platform orders; state-media source reporting company figures.", "url": "https://flyvolo.ai/en/changes/ev-20250422-delivery-rider-1" }, { "id": "ev-20250428-translator-1", "occupation_slug": "translator", "task_ids": "bulk-translation", "title": "Duolingo's CEO told staff the company would become 'AI-first' and gradually stop using contractors for work AI can handle; the memo was later softened after public backlash", "stage": "labor_impact", "occurred_on": "2025-04-28", "verified_on": "2026-09-13", "source_name": "Tech Wire Asia", "source_url": "https://techwireasia.com/2025/04/duolingo-shifts-toward-ai-first-model-phases-out-contractor-roles/", "source_tier": "secondary", "scope": "One company's stated policy for contractor work (content and translation). Primary source is a LinkedIn post by the company.", "url": "https://flyvolo.ai/en/changes/ev-20250428-translator-1" }, { "id": "ev-20250501-hr-recruiter-3", "occupation_slug": "hr-recruiter", "task_ids": "hr-administration", "title": "IBM reported its AskHR agent containing 94% of common HR questions across more than 80 automated HR tasks, with support tickets down 75% since 2016 and HR operating costs down 40% over four years", "stage": "labor_impact", "occurred_on": "2025-05-01", "verified_on": "2026-09-11", "source_name": "IBM — AskHR case study (first-party)", "source_url": "https://www.ibm.com/case-studies/ibm-askhr", "source_tier": "primary", "scope": "One employer, its own published figures, covering internal HR service delivery for a workforce of over 270,000 — vacation, payroll and policy queries, not hiring decisions or employee relations. This is a change in job scope and cost, not a headcount figure: IBM does not publish HR headcount here. The 75% ticket fall is measured from 2016, so it spans a decade of self-service automation, most of it predating generative AI.", "url": "https://flyvolo.ai/en/changes/ev-20250501-hr-recruiter-3" }, { "id": "ev-20250501-truck-driver-1", "occupation_slug": "truck-driver", "task_ids": "highway-driving", "title": "Aurora began regular driverless commercial freight deliveries between Dallas and Houston with no human on board, for Uber Freight and Hirschbach", "stage": "deployment", "occurred_on": "2025-05-01", "verified_on": "2026-09-10", "source_name": "Aurora Innovation — press release", "source_url": "https://ir.aurora.tech/news-events/press-releases/detail/119/aurora-begins-commercial-driverless-trucking-in-texas-ushering-in-a-new-era-of-freight", "source_tier": "primary", "scope": "US, Texas, one interstate corridor, heavy-duty trucks; hub-to-hub with human-handled first and last segments. Expansion to El Paso and Phoenix announced for end-2025.", "url": "https://flyvolo.ai/en/changes/ev-20250501-truck-driver-1" }, { "id": "ev-20250508-ai-implementation-lead-3", "occupation_slug": "ai-implementation-lead", "task_ids": "testing-on-our-own-work", "title": "China's labour ministry put 17 new occupations and 42 new work-types out for comment, adding a 'generative AI system tester' work-type under the existing 'generative AI system operator' occupation", "stage": "labor_impact", "occurred_on": "2025-05-08", "verified_on": "2026-09-11", "source_name": "中国就业网(人力资源和社会保障部)——《职业上新!42 个新工种亮相》,2025-05-09", "source_url": "https://chinajob.mohrss.gov.cn/c/2025-05-09/433983.shtml", "source_tier": "primary", "scope": "Mainland China, and the occupational classification only. The ministry's stated reason is that as large models spread across industries, 'the number of practitioners is growing rapidly and the range of posts is becoming more varied' — that is the authority of record saying this work now exists at classifiable scale, and it is the whole of what this record establishes. It gives no headcount, no wages and no employer names. It is a 公示 (public notice for comment): at this date the categories were proposed, not yet written into the national occupational catalogue, and a national occupational standard was still to be drafted.", "url": "https://flyvolo.ai/en/changes/ev-20250508-ai-implementation-lead-3" }, { "id": "ev-20250508-customer-service-representative-2", "occupation_slug": "customer-service-representative", "task_ids": "faq-answering; quality-oversight", "title": "Klarna's CEO said the company would again hire humans for customer service, saying cost had been 'too predominant' in its AI-first approach and quality had suffered", "stage": "constraint", "occurred_on": "2025-05-08", "verified_on": "2026-09-10", "source_name": "CX Dive (reporting a Bloomberg interview)", "source_url": "https://www.customerexperiencedive.com/news/klarna-reinvests-human-talent-customer-service-AI-chatbot/747586/", "source_tier": "secondary", "scope": "Same company as the February 2024 deployment; the assistant kept handling roughly two-thirds of inquiries. Secondary source quoting a Bloomberg interview of 8 May 2025.", "url": "https://flyvolo.ai/en/changes/ev-20250508-customer-service-representative-2" }, { "id": "ev-20250513-school-teacher-3", "occupation_slug": "school-teacher", "task_ids": "orchestrating-learning-tools; lesson-preparation", "title": "China's education ministry advisory committee issued an AI general-education guideline and a generative-AI use guideline for all primary and secondary schools", "stage": "mandate", "occurred_on": "2025-05-13", "verified_on": "2026-09-12", "source_name": "中华人民共和国教育部 科技与信息化司——《2025 年 5 月教育信息化和网络安全工作月报》", "source_url": "http://www.moe.gov.cn/s78/A16/gongzuo/gzzl_yb/202506/t20250630_1195963.html", "source_tier": "primary", "scope": "Mainland China, all primary and secondary schools, issued 12 May 2025. The ministry's own description: the education guideline builds a general AI education system centred on literacy, through a spiral curriculum running from early cognition to innovative practice; the use guideline covers generative-AI application scenarios in schools, sets usage norms for each stage, and holds data-security and ethical lines while the technology assists teaching and personalised learning. Two things this source does NOT establish, which earlier wording claimed: it does not describe a strengthened guidance role for teachers, and it does not set out the three named stages. It is a guideline, not an implementation record — it says what schools are to do, not what any school did, and there is no headcount, budget or timetable in it.", "url": "https://flyvolo.ai/en/changes/ev-20250513-school-teacher-3" }, { "id": "ev-20250518-journalist-1", "occupation_slug": "journalist", "task_ids": "commodity-rewrites; verification", "title": "The Chicago Sun-Times and Philadelphia Inquirer printed a syndicated summer reading list with books that do not exist; the freelancer said he used an AI tool and the section was not reviewed", "stage": "constraint", "occurred_on": "2025-05-18", "verified_on": "2026-09-10", "source_name": "NBC News", "source_url": "https://www.nbcnews.com/tech/tech-news/chicago-sun-admits-summer-book-guide-included-fake-ai-generated-titles-rcna208325", "source_tier": "secondary", "scope": "US newspapers; a syndicated special section produced by King Features (Hearst), not by newsroom staff. King Features said it ended its relationship with the writer; the paper's CEO called it a failure of review, not of reporting.", "url": "https://flyvolo.ai/en/changes/ev-20250518-journalist-1" }, { "id": "ev-20250520-financial-analyst-3", "occupation_slug": "financial-analyst", "task_ids": "model-building", "title": "On the Finance Agent Benchmark — 537 expert-authored questions over recent SEC filings — the best model, OpenAI o3, reached 46.8% accuracy at $3.79 per query", "stage": "constraint", "occurred_on": "2025-05-20", "verified_on": "2026-09-11", "source_name": "arXiv 2508.00828 (Bigeard, Nashold, Krishnan, Wu)", "source_url": "https://arxiv.org/abs/2508.00828", "source_tier": "primary", "scope": "Nine task categories from information retrieval to financial modelling, authored with experts from banks, hedge funds and private equity; the agents were given Google Search and EDGAR access. Public-filing research only — not an analyst working inside their own firm's models and internal data.", "url": "https://flyvolo.ai/en/changes/ev-20250520-financial-analyst-3" }, { "id": "ev-20250520-pharmacist-1", "occupation_slug": "pharmacist", "task_ids": "dispensing; clinical-services", "title": "Walgreens opened its 12th robotic micro-fulfilment centre; the network fills 3.5 million+ prescriptions a week for 5,000+ stores, about 40% of a supported store's prescription volume", "stage": "deployment", "occurred_on": "2025-05-20", "verified_on": "2026-09-10", "source_name": "Walgreens — press release", "source_url": "https://corporate.walgreens.com/newsroom/walgreens-expands-micro-fulfillment-network-with-new-facility-in-brooklyn-park-minnesota/", "source_tier": "primary", "scope": "One US chain; central fill of a share of routine prescriptions, the rest stays in-store. Company-reported; the release says freed pharmacy time goes to vaccinations and adherence support. Trade press notes this was the first opening since the chain paused expansion in autumn 2024.", "url": "https://flyvolo.ai/en/changes/ev-20250520-pharmacist-1" }, { "id": "ev-20250601-counsellor-2", "occupation_slug": "counsellor", "task_ids": "making-a-therapeutic-decision", "title": "Japan's certified psychologist law requires the practitioner to take the instruction of the attending physician where the person has one, and makes using the protected title a criminal offence", "stage": "constraint", "occurred_on": "2025-06-01", "verified_on": "2026-09-13", "source_name": "e-Gov 法令検索 (Japan, Digital Agency) — 公認心理師法 第四十二条・第四十四条・第四十九条", "source_url": "https://laws.e-gov.go.jp/law/427AC0000000068", "source_tier": "primary", "scope": "Read on the Digital Agency's own statute portal, in the version in force from 1 June 2025; the Act dates from 2015. Two provisions bear on this page and they work differently. Article 42(2) is a duty on the practitioner: where a person needing psychological support has an attending physician for that support, the certified psychologist shall take that physician's instruction. Article 42(1) separately requires the practitioner to maintain coordination with those providing health, welfare and education services. Article 44 protects the title — a person who is not certified may not call themselves 公認心理師, nor use the characters 心理師 within any name they use — and article 49 makes breaching that punishable by a fine of up to ¥300,000, so the title restriction is criminal rather than professional. What this does not do is regulate any technology: the Act predates the services this page is about and names none of them. It constrains who may hold themselves out under this title and whose instruction binds a therapeutic decision, which is a fact about the professional structure rather than about what software can do. It applies in Japan and counts nobody.", "url": "https://flyvolo.ai/en/changes/ev-20250601-counsellor-2" }, { "id": "ev-20250601-cybersecurity-analyst-4", "occupation_slug": "cybersecurity-analyst", "task_ids": "hunting", "title": "The US Justice Department reported a user-behaviour monitoring system in production since June 2025 that it says finds behaviour too subtle for a person to detect", "stage": "deployment", "occurred_on": "2025-06-01", "verified_on": "2026-09-20", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory (entry DOJ-0119)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory", "source_tier": "primary", "scope": "One bureau (EOUSA) inside one department, reported under a statutory duty rather than chosen for publication. Entry DOJ-0119 in the individually-reported CSV. The inventory gives the month, not the day, so the date here is the first of that month. The claim that matters is the department describing the problem it solves as \"identification and prevention of anomalous user behavior too subtle for human detection and analysis\" — an employer stating in a filing that a machine covers ground its people cannot. Two things are in the same rows and belong beside it. The department classified this system High-impact, and every safeguard the high-impact classification requires is recorded as in progress: testing, impact assessment, independent review, ongoing monitoring, training, failsafe, appeal process and public consultation, with authority to operate answered No. The system holds personal data. Separately, entry GSA \"Elastic Machine Learning Threat Detection (Phase 2)\" describes the other side of the same boundary in a different agency, saying its models allow \"isolation and ranking of cyber threats, facilitating faster and more accurate escalation to threat hunting and incident response teams\" — the machine ranks and the hunting team still receives. Neither entry reports a headcount, a caseload or a measured detection rate, and the department's own evaluation of whether the claim holds is one of the items marked in progress.", "url": "https://flyvolo.ai/en/changes/ev-20250601-cybersecurity-analyst-4" }, { "id": "ev-20250601-physical-therapist-2", "occupation_slug": "physical-therapist", "task_ids": "the-programme; hands-on-treatment", "title": "Japan defines a physical therapist as someone licensed to practise under a physician's direction, and treats the therapy itself as assistance to medical care", "stage": "constraint", "occurred_on": "2025-06-01", "verified_on": "2026-09-13", "source_name": "e-Gov 法令検索 (Japan, Digital Agency) — 理学療法士及び作業療法士法 第二条・第十五条", "source_url": "https://laws.e-gov.go.jp/law/340AC0000000137", "source_tier": "primary", "scope": "Read on the Digital Agency's own statute portal, in the version in force from 1 June 2025; the Act dates from 1965. The constraint is in the definition rather than in a rule bolted on afterwards: Article 2(3) defines a physical therapist as a person licensed by the Minister who, using that title, carries on physical therapy as a business under a physician's direction. Article 15(1) then permits physical and occupational therapists to practise, notwithstanding the Nurses Act, as assistance to medical care — the legal category their work sits in. Article 15(2) is narrower and more specific: massage performed as physical therapy escapes the separate licensing law for masseurs only where it is done in a hospital or clinic, or under a physician's concrete instruction. So the physician's direction is general for the practice and specific for that hands-on element outside an institution. What this does not do is say who or what may prepare the assessment, propose the programme or track the progress; the Act governs the authority under which treatment is given, not the method by which it is worked out. It binds practice in Japan and counts nobody.", "url": "https://flyvolo.ai/en/changes/ev-20250601-physical-therapist-2" }, { "id": "ev-20250601-security-guard-2", "occupation_slug": "security-guard", "task_ids": "walking-toward-it; watching-screens", "title": "Japan's Security Services Act lets a base station do the monitoring but requires operators to keep enough guards, standby posts and vehicles deployed for a guard to confirm the facts at the scene", "stage": "constraint", "occurred_on": "2025-06-01", "verified_on": "2026-09-13", "source_name": "e-Gov 法令検索 (Japan, Digital Agency) — 警備業法 第四十三条(即応体制の整備)・第四十四条", "source_url": "https://laws.e-gov.go.jp/law/347AC0000000117", "source_tier": "primary", "scope": "Read on the Digital Agency's own statute portal, in the version in force from 1 June 2025; the Act dates from 1972. The structure is the useful part, because it splits this occupation's work in two and treats each half differently. The Act defines machine security services as a regulated business in their own right — sensors at the protected site transmitting incident information to equipment somewhere else — so remote electronic monitoring is not a workaround, it is a licensed category with its own registration. Article 43 then attaches the condition: an operator of such a service must, following standards set by the prefectural public safety commission, keep properly deployed the necessary number of guards, standby premises where those guards wait, and vehicles and other equipment, so that when the base station receives information about a theft or other incident, confirmation of the facts by a guard at the scene and other necessary measures can be taken promptly. Article 44 makes it individual rather than notional: each base station must keep a document naming the guards assigned to each standby post. What this does not do is set a national response time — the standards are left to each prefecture's commission, so the law does not itself say how fast. It binds operators in Japan, counts nobody, and says nothing about how many sites are monitored this way.", "url": "https://flyvolo.ai/en/changes/ev-20250601-security-guard-2" }, { "id": "ev-20250623-experienced-software-engineer-2", "occupation_slug": "experienced-software-engineer", "task_ids": "reviewing-what-the-machine-wrote", "title": "Stack Overflow's survey of 49,009 developers finds 84% using or planning to use AI tools, and the most experienced reporting the highest distrust of what it produces", "stage": "worker_adoption", "occurred_on": "2025-06-23", "verified_on": "2026-09-15", "source_name": "Stack Overflow 2025 Developer Survey — AI section(平台自己的年度普查)", "source_url": "https://survey.stackoverflow.co/2025/ai", "source_tier": "primary", "scope": "49,009 responses from 177 countries, fielded 29 May to 23 June 2025, and self-selected: respondents were recruited through Stack Overflow's own channels, so this measures the people who answer Stack Overflow, not a random sample of developers. It is recorded against reviewing machine-written code rather than against writing code, because what it measures is the volume driver and the verification response, not who writes what. Both halves belong here: 84% use or plan to use AI tools (76% a year earlier) and 51% of professional developers use them daily, while favourable sentiment fell from over 70% in 2023 and 2024 to 60%, trust in accuracy dropped to 29% among professional developers, and the most experienced report the lowest rate of high trust at 2.6% against the highest rate of high distrust at 20%. The stake cuts both ways and that is unusual enough to state: Stack Overflow's own traffic is what AI coding assistants displace, so the adoption finding runs against its commercial interest while the falling-trust finding runs with it. The survey does not say what the tools were used for, whether any review took longer, or what any of it changed about shipped code.", "url": "https://flyvolo.ai/en/changes/ev-20250623-experienced-software-engineer-2" }, { "id": "ev-20250623-financial-analyst-1", "occupation_slug": "financial-analyst", "task_ids": "data-gathering-and-summaries", "title": "Goldman Sachs made its in-house GS AI Assistant available to all employees after use by thousands of staff, for summarising complex documents, drafting initial content and data analysis", "stage": "deployment", "occurred_on": "2025-06-23", "verified_on": "2026-09-10", "source_name": "Fox Business (reporting a Reuters-obtained internal memo)", "source_url": "https://www.foxbusiness.com/technology/goldman-sachs-announces-firmwide-launch-ai-assistant", "source_tier": "secondary", "scope": "One US investment bank; a firmwide general-purpose assistant with versions for bankers, research analysts and wealth staff, announced in a memo by CIO Marco Argenti. Reported from the internal memo (Reuters first); no headcount statement.", "url": "https://flyvolo.ai/en/changes/ev-20250623-financial-analyst-1" }, { "id": "ev-20250630-content-moderator-2", "occupation_slug": "content-moderator", "task_ids": "the-appeal", "title": "Meta reports 2,253,559 content-removal appeals on Facebook in the EU in six months, 786,034 restored, and states that all Article 16 notices are processed using manual review", "stage": "constraint", "occurred_on": "2025-06-30", "verified_on": "2026-09-22", "source_name": "Meta Platforms Ireland, DSA Transparency Report for Facebook", "source_url": "https://transparency.meta.com/sr/dsa-transparency-report-aug2025-facebook", "source_tier": "primary", "scope": "The platform's own statutory transparency report under the Digital Services Act, covering 1 January to 30 June 2025, published 29 August 2025. Three things in it, read from the tables rather than from any summary. Table 15.1.d.(1): 2,253,559 organic content-removal complaints and 786,034 restorations after complaint, which is 34.9% - and the spread by category is wide, from 41,393 complaints with 38,730 restorations under Cybersecurity to 197,216 with 22,804 under Hateful Conduct. Stat 15.1.b.(3), whose heading is Notices processed by using automated means, answers that all Article 16 DSA notices are processed using manual review. And the report states that Meta employs a combination of human review and technology, with most own-initiative removals happening automatically. Put beside the same publisher's 94.1% automated-action figure, that is the finding: one channel at this platform is almost entirely machine and the statutory notice channel next to it is entirely human. What the report does not say is who decides an appeal - it gives the volumes and the outcomes and never states whether the second look is taken by a person or by another system. That silence is the reason this record does not settle the direction of the task by itself.", "url": "https://flyvolo.ai/en/changes/ev-20250630-content-moderator-2" }, { "id": "ev-20250701-auditor-3", "occupation_slug": "auditor", "task_ids": "sampling-and-testing", "title": "PCAOB economists reported registered audit firm staffing rising every year from 2015 to 2024, and rising per dollar of client revenue, with non-CPA accountants growing about twice as fast as CPAs", "stage": "labor_impact", "occurred_on": "2025-07-01", "verified_on": "2026-09-12", "source_name": "PCAOB Office of Economic and Risk Analysis — Data Points: Registered Firm Staffing Trends", "source_url": "https://assets.pcaobus.org/pcaob-dev/docs/default-source/economicandriskanalysis/data-points/2025-dp01_registeredfirmstaffingtrends.pdf?sfvrsn=383aa3eb_1", "source_tier": "primary", "scope": "US registered public accounting firms only, counted from the staffing figures those firms file with the regulator on Form 2, against issuer revenues from Compustat as a proxy for audit volume. The second measure is the one that matters here: staffing rose even after dividing by the revenue of the companies being audited, so the rise is not simply clients getting bigger. Two limits the report states itself — registered firm staff may work on matters unrelated to issuer audits, and firms whose issuer clients had no Compustat revenue were excluded. The period ends in 2024, and headcount is the crudest of the measures that could move: it says nothing about what those people spend the day doing. The composition shift is worth reading separately from the total: the fast-growing half is non-CPA accountants, who cannot sign.", "url": "https://flyvolo.ai/en/changes/ev-20250701-auditor-3" }, { "id": "ev-20250701-translator-2", "occupation_slug": "translator", "task_ids": "bulk-translation; post-editing", "title": "Crunchyroll's German subtitles for an anime premiere contained the line 'ChatGPT said…'; the company said a third-party vendor had used AI-generated subtitles in violation of its agreement", "stage": "constraint", "occurred_on": "2025-07-01", "verified_on": "2026-09-10", "source_name": "Engadget", "source_url": "https://www.engadget.com/entertainment/streaming/crunchyroll-blames-third-party-vendor-for-ai-subtitle-mess-145621606.html", "source_tier": "secondary", "scope": "Entertainment subtitling, one streaming platform, one episode (Necronomico and the Cosmic Horror Show, ep. 1, 1 July 2025). The company's president had said in 2024 it was testing generative AI for subtitling; the failure was unreviewed vendor output.", "url": "https://flyvolo.ai/en/changes/ev-20250701-translator-2" }, { "id": "ev-20250701-warehouse-worker-1", "occupation_slug": "warehouse-worker", "task_ids": "goods-movement; robot-fleet-operation", "title": "Amazon deployed its one-millionth robot across 300+ facilities; its Shreveport next-generation site needs 30% more staff in reliability, maintenance and engineering roles", "stage": "deployment", "occurred_on": "2025-07-01", "verified_on": "2026-09-10", "source_name": "Amazon — About Amazon", "source_url": "https://www.aboutamazon.com/news/operations/amazon-million-robots-ai-foundation-model", "source_tier": "primary", "scope": "Amazon's global fulfilment network; company-reported. The robot count is fleet size, not a statement about total headcount.", "url": "https://flyvolo.ai/en/changes/ev-20250701-warehouse-worker-1" }, { "id": "ev-20250710-junior-software-developer-2", "occupation_slug": "experienced-software-engineer", "task_ids": "changing-a-system-you-did-not-write", "title": "A METR randomised trial of 16 experienced open-source developers on 246 real issues found they took 19% longer with early-2025 AI tools, while believing they had been 20% faster", "stage": "constraint", "occurred_on": "2025-07-10", "verified_on": "2026-09-12", "source_name": "METR — study report", "source_url": "https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/", "source_tier": "primary", "scope": "Experienced maintainers on large, mature repositories they know well (about 5 years each); tools were mainly Cursor Pro with Claude 3.5/3.7 Sonnet. The authors explicitly do not claim the result generalises to most developers or to greenfield or junior work.", "url": "https://flyvolo.ai/en/changes/ev-20250710-junior-software-developer-2" }, { "id": "ev-20250722-journalist-2", "occupation_slug": "journalist", "task_ids": "commodity-rewrites", "title": "Pew found Google users clicked a search result in 8% of visits when an AI summary appeared, against 15% when it did not, and clicked a link inside the summary in 1%", "stage": "deployment", "occurred_on": "2025-07-22", "verified_on": "2026-09-11", "source_name": "Pew Research Center", "source_url": "https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results", "source_tier": "primary", "scope": "United States, March 2025, 900 adults who agreed to share their browsing — a behavioural panel, not the whole web, and Google only. It measures clicking, not revenue and not employment. One finding cuts against the simple reading: news sites made up 5% of the sources cited in AI summaries and also 5% of those in standard results, so the summaries are not citing news less than ordinary search did. What collapsed is the click, not the citation.", "url": "https://flyvolo.ai/en/changes/ev-20250722-journalist-2" }, { "id": "ev-20250801-counsellor-1", "occupation_slug": "counsellor", "task_ids": "making-a-therapeutic-decision", "title": "Illinois enacted the Wellness and Oversight for Psychological Resources Act, barring AI from mental health and therapeutic decision-making and permitting it only for administrative support", "stage": "constraint", "occurred_on": "2025-08-01", "verified_on": "2026-09-12", "source_name": "Illinois Department of Financial and Professional Regulation (state agency announcement)", "source_url": "https://idfpr.illinois.gov/news/2025/gov-pritzker-signs-state-leg-prohibiting-ai-therapy-in-il.html", "source_tier": "primary", "scope": "One US state, announced by the state agency that regulates the professions concerned. It is regulation arriving ahead of deployment rather than after it, which is unusual on this site. It does not bind other markets, and it does not reach products marketed as wellness or coaching rather than therapy — which is where most consumer products currently sit. A law also states what may not be offered, not what people actually use.", "url": "https://flyvolo.ai/en/changes/ev-20250801-counsellor-1" }, { "id": "ev-20250806-actuary-2", "occupation_slug": "actuary", "task_ids": "the-price; the-reserves; fairness-testing; the-opinion", "title": "EIOPA's Opinion made the actuarial function responsible for the controls on AI systems within its remit, and recorded that AI for life and health insurance pricing is high-risk under the AI Act", "stage": "constraint", "occurred_on": "2025-08-06", "verified_on": "2026-09-23", "source_name": "European Insurance and Occupational Pensions Authority — Opinion on Artificial Intelligence governance and risk management, EIOPA-BoS-25-360", "source_url": "https://www.eiopa.europa.eu/publications/opinion-artificial-intelligence-governance-and-risk-management_en", "source_tier": "primary", "scope": "The European insurance supervisor's own Opinion of 6 August 2025, addressed to national competent authorities rather than directly to firms. Read in the full PDF. Paragraph 2.4: the AI Act identifies as high-risk the use of AI systems for risk assessment and pricing in relation to natural persons in life and health insurance. In its description of roles, it states that the actuarial function is responsible for the controls on AI systems that fall under its responsibilities, giving as examples the coordination of the technical provisions calculation and the opinion on the overall underwriting policy. Its annex on fairness metrics warns that some group fairness metrics could contradict actuarial fairness, where customers bearing the same risk are charged the same price. It establishes where responsibility sits in EU insurers; it does not measure how much of actuarial work software does, and it counts no one.", "url": "https://flyvolo.ai/en/changes/ev-20250806-actuary-2" }, { "id": "ev-20250806-insurance-claims-handler-2", "occupation_slug": "insurance-claims-handler", "task_ids": "deciding-whether-it-is-covered; telling-someone-no", "title": "EIOPA's Opinion requires EU insurers to keep effective internal control and human oversight over AI systems, and records that claims AI is not on the AI Act's high-risk list", "stage": "constraint", "occurred_on": "2025-08-06", "verified_on": "2026-09-12", "source_name": "European Insurance and Occupational Pensions Authority — Opinion on Artificial Intelligence governance and risk management, EIOPA-BoS-25-360", "source_url": "https://www.eiopa.europa.eu/publications/opinion-artificial-intelligence-governance-and-risk-management_en", "source_tier": "primary", "scope": "European Union, addressed to national competent authorities rather than directly to firms, and covering insurers and intermediaries across the value chain. Two things have to be read together or the record misleads. What it requires: following Solvency II Article 46 and Article 258 of Delegated Regulation 2015/35, effective internal control across the whole AI lifecycle, with roles, escalation procedures and named responsible functions written into policy, staff trained for their role, and guardrails so the system behaves as intended. What it is not: paragraph 2.8 says the Opinion sets out no new requirements, and paragraphs 2.4 to 2.6 record that the AI Act's high-risk list reaches risk assessment and pricing in life and health insurance — not claims handling. So automating a claims decision in the EU faces no high-risk regime; what applies is ordinary internal-control law, plus a duty to tell the customer they are dealing with an AI system and to promote staff AI literacy. Nothing here requires a person to decide any individual claim.", "url": "https://flyvolo.ai/en/changes/ev-20250806-insurance-claims-handler-2" }, { "id": "ev-20250831-management-consultant-2", "occupation_slug": "management-consultant", "task_ids": "the-deck", "title": "Accenture's annual report to August 2025 states a talent strategy whose second prong is exiting people where reskilling is not viable, and books $344 million of severance against it", "stage": "labor_impact", "occurred_on": "2025-08-31", "verified_on": "2026-09-22", "source_name": "Accenture plc — Form 10-K for the fiscal year ended 31 August 2025 (filed 2025-10-10)", "source_url": "https://www.sec.gov/Archives/edgar/data/1467373/000146737325000217/acn-20250831.htm", "source_tier": "primary", "scope": "The source is the company's own annual report on Form 10-K, filed with the SEC on 10 October 2025 for the fiscal year ended 31 August 2025, so this is the employer describing its own workforce rather than a consultancy describing someone else's. The filing sets out three prongs: investing in upskilling, which it calls its primary focus; exiting people in a compressed timeline where reskilling is not a viable path for the skills it needs; and identifying areas to drive operating efficiencies, including through AI. Read the order carefully, because it is the whole finding: AI appears in the third prong and the filing does not say AI caused the exits. The same document carries the counterweight — approximately $1.0 billion and approximately 47 million training hours spent on learning in the same year, and approximately 779,000 people employed at the year end. The $344 million of severance sits inside $615 million of business optimisation costs whose remainder is asset impairments from two divestitures. The task link is our judgement, not the filing's. The document itself names no task. Our practice is to link a firm-level workforce record to the task the effect most directly concerns. One firm is not the occupation, and US employment in management consulting rose over the same period.", "url": "https://flyvolo.ai/en/changes/ev-20250831-management-consultant-2" }, { "id": "ev-20250901-radiologist-5", "occupation_slug": "radiologist", "task_ids": "answering-for-the-machine-that-read-it; the-report-somebody-acts-on", "title": "The Royal College of Radiologists' 2025 census, answered by every UK radiology department, found AI tools in 75% of them, and no workload reduction for over three-quarters of clinical directors", "stage": "deployment", "occurred_on": "2025-09-01", "verified_on": "2026-09-20", "source_name": "The Royal College of Radiologists — Clinical Radiology Census 2025", "source_url": "https://www.rcr.ac.uk/media/n1fjvrv4/rcr-2025-clinical-radiology-workforce-census-report.pdf", "source_tier": "primary", "scope": "The United Kingdom only, and hospital radiology departments only. This is the College's 18th annual census of its own members, answered by clinical directors with a 100% completion rate, so its adoption figures are a count of a defined population rather than a sample. Two clauses carry the linked tasks. On answering for the machine, the report states that additional time would also be required by consultants to oversee and monitor the AI system itself, and prints a clinical radiologist's own words: it slows down my acute stroke reporting as I have to check the AI and I often conclude that it is incorrect, I then have to explain this in my report. On the report itself, AI tools to aid in drafting reports are in place in 11% of departments and, of those using them, 24% of clinical directors said the tools reduced their workload, which is the most positive figure for any use case in this census and still a minority. Overall adoption: 75% of departments use AI tools in some capacity and support for image interpretation is in place in 58%. Counter-signals from the same document, which are the reason it is worth recording: across every use case more than three-quarters of clinical directors report that AI tools are either having no impact or are increasing their workloads; AI implementation ranked last of seven productivity initiatives at 28% positive against 70% for more skill mix and administrative support; and the report concludes that at this stage it does not appear that AI tools in radiology are simultaneously enabling faster and more accurate reporting, detection of more disease, and a reduction in the backlog. What it does not establish: it counts departments that hold a tool, not studies a machine read, so it says nothing about what share of reading is machine-first; and the workload figures are clinical directors' judgements rather than measured time, while the adoption figures are counts. The College has a stake and it runs both ways, since it is campaigning for more training places and cites these findings in support, and it also recommends further government investment in AI tools. The date is derived from the census's own line that data collection began in September 2025; no collection end date is given and the report's copyright line reads June 2026.", "url": "https://flyvolo.ai/en/changes/ev-20250901-radiologist-5" }, { "id": "ev-20250910-civil-engineer-1", "occupation_slug": "civil-engineer", "task_ids": "the-model; the-submission", "title": "Singapore's URA confirmed CORENET X digital building submission becomes mandatory in phases, reaching all new projects from October 2026", "stage": "mandate", "occurred_on": "2025-09-10", "verified_on": "2026-09-23", "source_name": "Urban Redevelopment Authority — Circular URA/PB/2025/07-DCG, CORENET X: mandatory submissions starting from 1 October 2025", "source_url": "https://www.ura.gov.sg/guidelines/circulars/dc25-07/", "source_tier": "primary", "scope": "The Urban Redevelopment Authority's own circular of 10 September 2025, addressed to building owners, developers, architects, engineers, registered surveyors and contractors. It describes CORENET X as a one-stop digital shopfront for approval of building works that streamlines over 20 approval touchpoints across seven regulatory agencies into three gateways, with submissions in the openBIM format IFC+SG. Its timeline: mandatory for new projects of 30,000 square metres gross floor area or more from 1 October 2025, for all new projects regardless of size from 1 October 2026, and onboarding of all ongoing projects from 1 October 2027. It requires a digital route for submission; it does not measure time saved, it does not say who inside a firm builds the model, and it does not change who is responsible for the design.", "url": "https://flyvolo.ai/en/changes/ev-20250910-civil-engineer-1" }, { "id": "ev-20250925-assembly-line-worker-1", "occupation_slug": "assembly-line-worker", "task_ids": "repetitive-assembly", "title": "IFR's World Robotics 2025 counted 542,000 industrial robots installed in 2024 and 4.66 million in operation worldwide (+9%); Asia took 74% of new installations and China alone 54%", "stage": "deployment", "occurred_on": "2025-09-25", "verified_on": "2026-09-10", "source_name": "International Federation of Robotics — press release", "source_url": "https://ifr.org/ifr-press-releases/news/global-robot-demand-in-factories-doubles-over-10-years", "source_tier": "primary", "scope": "Global industry statistics from the robot manufacturers' federation, based on member and national-association data. Robot stock, not job counts; automotive and electronics dominate the installations.", "url": "https://flyvolo.ai/en/changes/ev-20250925-assembly-line-worker-1" }, { "id": "ev-20250925-assembly-line-worker-3", "occupation_slug": "assembly-line-worker", "task_ids": "repetitive-assembly", "title": "IFR counted 2,027,000 industrial robots working in Chinese factories, with 295,000 installed in 2024 — 54% of global demand and the highest annual figure ever recorded", "stage": "deployment", "occurred_on": "2025-09-25", "verified_on": "2026-09-11", "source_name": "International Federation of Robotics — World Robotics 2025, China press release (first-party)", "source_url": "https://ifr.org/downloads/press_docs/2025-09-25-IFR_press_release_China_in_English.pdf", "source_tier": "primary", "scope": "China, from the industry federation's own annual statistics. The stock doubled in three years, passing 1 million in 2021. Installations are concentrated: electrical and electronics took 83,000 units (64% of that industry's global installations), metal and machinery rose 31% to 54,600, while automotive fell 12% to 57,200. The labour-intensive sectors are the striking part — textiles, leather and apparel installed 5,700 units and China accounted for 95% of global installations in that category. Domestic suppliers outsold foreign ones at home for the first time, 57% in 2024 against 47% in 2023. These are robot counts, not headcount: the release says nothing about how many line workers those factories employ.", "url": "https://flyvolo.ai/en/changes/ev-20250925-assembly-line-worker-3" }, { "id": "ev-20250930-air-traffic-controller-1", "occupation_slug": "air-traffic-controller", "task_ids": "staffing-the-watch", "title": "FAA reports a net gain of 568 controllers after hiring 2,028 trainees in FY2025, and states it uses no automated scheduling optimisation tools", "stage": "labor_impact", "occurred_on": "2025-09-30", "verified_on": "2026-09-22", "source_name": "Federal Aviation Administration", "source_url": "https://www.faa.gov/sites/faa.gov/files/Air-Traffic-Controller-Workforce-Plan-2026-2028_0.pdf", "source_tier": "primary", "scope": "The agency's own Controller Workforce Plan for 2026-2028. Two separate things in one document. The workforce figures: 2,028 trainees hired in FY2025 against a target of 2,000 and the most since 2008, Academy starting salaries raised by close to thirty percent, total workforce losses of 1,460 over the same year, nearly 400 retirement-eligible controllers retained by a bonus, and a resulting net gain of 568, with targets of 2,200, 2,300 and 2,400 for the three following years. Those figures describe one country's provider and a headcount, not a task. The sentence that bears on the linked task is separate and unusually direct: the agency writes that it does not use any automated scheduling optimisation tools, that workforce scheduling and controller timekeeping are both accomplished manually by local facility managers, and that it is difficult to understand why none have been deployed. Note what that does and does not establish. It establishes an adoption gap at one employer for one task, in the employer's own words, in a document it is required to publish. It does not establish anything about other providers, and the same plan lists optimising scheduling efficiency as one of its three strategic pillars, so the practice it describes is one its own author is arguing against. The document carries no publication line of its own; the date recorded here is the last-modified date the agency's own server reports for the file.", "url": "https://flyvolo.ai/en/changes/ev-20250930-air-traffic-controller-1" }, { "id": "ev-20251002-bus-driver-1", "occupation_slug": "bus-driver", "task_ids": "the-drive; safety-operator; remote-supervision; getting-people-on", "title": "Singapore's Land Transport Authority awarded a contract to pilot six autonomous buses on two public bus services, with a safety operator on board at first and a remote operator later", "stage": "pilot", "occurred_on": "2025-10-02", "verified_on": "2026-09-23", "source_name": "Land Transport Authority (Singapore) — news release", "source_url": "https://www.lta.gov.sg/content/ltagov/en/newsroom/2025/10/news-releases/lta_awards_contract_pilot_deployment_autonomous_buses.html", "source_tier": "primary", "scope": "The authority's own release on services 400 and 191: six 16-seat buses, an initial three-year pilot from the second half of 2026, running alongside manned buses. It sets out the task split in two phases: a safety operator on board at all times first; then a remote operator monitoring continuously from a control centre, with a customer service officer on board to assist commuters. Existing bus captains are to be trained as safety operators, and the authority, operators and the transport workers' union will prepare training for new roles. It establishes a plan and a procurement, not an outcome: no bus had carried a passenger in this pilot when it was published, and it says nothing about how many drivers' jobs a network would need.", "url": "https://flyvolo.ai/en/changes/ev-20251002-bus-driver-1" }, { "id": "ev-20251005-procurement-specialist-1", "occupation_slug": "procurement-specialist", "task_ids": "finding-and-comparing; the-order-and-the-paperwork", "title": "OpenAI's GDPval benchmark collected real purchasing-agent deliverables and had industry experts grade model output against them blind", "stage": "capability", "occurred_on": "2025-10-05", "verified_on": "2026-09-12", "source_name": "GDPval: Evaluating AI Model Performance on Real-World Economically Valuable Tasks (OpenAI, arXiv:2510.04374)", "source_url": "https://arxiv.org/pdf/2510.04374", "source_tier": "primary", "scope": "Buyers and purchasing agents are one of 44 US occupations in the benchmark, across the nine sectors contributing most to US GDP; 1,320 tasks in the full set with at least 30 per occupation, each built from actual work product by a professional averaging 14 years of experience and covering the majority of that occupation's O*NET work activities. What is graded is a deliverable, one shot, by a blind pairwise expert comparison - not a job. There is no client, no revision round, no negotiation and no consequence for being wrong. The headline figure (47.6% of one model's gold-subset deliverables rated better than or as good as the expert's) is for the model generation of September 2025 and has been superseded; more usefully, the paper reports that win rates are highest on tasks taking 0-2 hours and decline steadily as the task gets longer. Per-occupation rates are published only as a figure, so no purchasing-specific number is quoted here. OpenAI designed the benchmark, ran it, and sells models measured by it.", "url": "https://flyvolo.ai/en/changes/ev-20251005-procurement-specialist-1" }, { "id": "ev-20251007-accountant-2", "occupation_slug": "auditor", "task_ids": "owning-machine-drafted-work", "title": "Deloitte agreed to repay the last instalment of a US$290,000 Australian government report after fabricated references and a misattributed court quote were found; it acknowledged using Azure OpenAI", "stage": "constraint", "occurred_on": "2025-10-07", "verified_on": "2026-09-13", "source_name": "Fortune", "source_url": "https://fortune.com/2025/10/07/deloitte-ai-australia-government-report-hallucinations-technology-290000-refund", "source_tier": "secondary", "scope": "An assurance review for Australia's Department of Employment and Workplace Relations by a Big Four firm's consulting practice — professional-services output, not bookkeeping. Corrected report issued 26 Sep 2025; the repaid instalment was later put at about A$97,000 (CFO Dive, 21 Oct 2025). Bears on who owns machine-drafted output, not on accounting tasks directly.", "url": "https://flyvolo.ai/en/changes/ev-20251007-accountant-2" }, { "id": "ev-20251008-counsellor-3", "occupation_slug": "counsellor", "task_ids": "being-available-at-three-in-the-morning", "title": "NHS England published its assessments of digitally delivered therapy products, recording three of eleven as not compliant and three more as approved only when a therapist routinely delivers them", "stage": "constraint", "occurred_on": "2025-10-08", "verified_on": "2026-09-20", "source_name": "NHS England — Digitally enabled therapies assessment criteria", "source_url": "https://www.england.nhs.uk/mental-health/adults/nhs-talking-therapies/digital/assessment-criteria/", "source_tier": "primary", "scope": "England only, and only products entering NHS Talking Therapies for anxiety and depression — private apps and other health systems are outside it. What it establishes is a gate, not a deployment: the page says the assessment \"gives services more confidence in the technologies they consider, select, commission and use locally\", so local services decide. Read the conditions rather than the counts: Koa Health Perspectives was recorded as not compliant because it \"is designed to deliver a low-intensity equivalent treatment\" for a condition its low-intensity practitioners are not trained in, and three 2025 products are compliant only \"provided the product is routinely delivered with the support of a high-intensity CBT therapist\". Of the chatbot product assessed, Wysa for generalised anxiety, NHS England wrote that it is \"compliant with the core and clinical content criteria\" but that \"there is not yet sufficient research demonstrating impact\". The page carries no publication date of its own, and NHS England does not publish the individual assessment dates; the date here is the earliest Internet Archive capture holding the 2025 batch (2025-10-08), and the previous capture, 2025-06-15, does not, so the outcomes became public between those two dates. The page also still says \"Six products have now been assessed\" while listing eleven.", "url": "https://flyvolo.ai/en/changes/ev-20251008-counsellor-3" }, { "id": "ev-20251011-registered-nurse-3", "occupation_slug": "registered-nurse", "task_ids": "patient-and-family-communication", "title": "California enacted AB 489, barring AI systems from using titles such as 'nurse' or 'doctor' where this implies a licensed person is providing the care", "stage": "constraint", "occurred_on": "2025-10-11", "verified_on": "2026-09-11", "source_name": "California Legislative Information (AB 489, chaptered)", "source_url": "https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260AB489", "source_tier": "primary", "scope": "California only, and only patient-facing AI that implies licensure. It does not restrict AI that discloses it is AI, nor internal clinical tools. Enforcement is by the licensing boards under Division 2 of the Business and Professions Code; each prohibited use is a separate violation.", "url": "https://flyvolo.ai/en/changes/ev-20251011-registered-nurse-3" }, { "id": "ev-20251016-ecommerce-operator-1", "occupation_slug": "ecommerce-operator", "task_ids": "the-listing; buying-traffic", "title": "Alibaba said its 2025 11.11 was its first large-scale deployment of generative AI in e-commerce, with an AI bidding engine it reports lifting marketing return by 12%", "stage": "deployment", "occurred_on": "2025-10-16", "verified_on": "2026-09-13", "source_name": "Alibaba Group (company announcement, 11.11 2025)", "source_url": "https://www.alibabagroup.com/document-1915930722120499200", "source_tier": "primary", "scope": "Alibaba's own announcement about its own platforms, so every figure is the platform's and none is independently checkable — the 12% marketing-return uplift, the roughly 200 million images and 5 million videos the merchant toolkit generates monthly, and the customer-service cost saving are all self-reported by the party selling the tools. Recorded anyway because what it establishes is not the size of the benefit but the fact of the deployment, and that fact is what this occupation's page turns on: the listing tools and the bidding engine are platform features, so a merchant does not choose whether to adopt them. It describes Chinese marketplace platforms at one shopping festival and does not generalise to other markets or to a brand running its own site.", "url": "https://flyvolo.ai/en/changes/ev-20251016-ecommerce-operator-1" }, { "id": "ev-20251022-devops-engineer-1", "occupation_slug": "devops-engineer", "task_ids": "writing-the-config; the-pager", "title": "DORA's 2025 survey reports 90% of respondents using AI at work, and finds AI adoption still has a negative relationship with software delivery stability", "stage": "worker_adoption", "occurred_on": "2025-10-22", "verified_on": "2026-09-13", "source_name": "Google Cloud, announcing the 2025 DORA report (the DORA research programme's own publication)", "source_url": "https://cloud.google.com/blog/products/ai-machine-learning/announcing-the-2025-dora-report", "source_tier": "primary", "scope": "A practitioner survey, so the 90% is self-reported use rather than measured use, and the accompanying '80% believe it increased their productivity' is a belief and must be read as one — this site holds a randomised trial in which experienced developers were 19% slower while believing they were 20% faster, so belief about one's own speed is not evidence about speed. What is not self-report about one's own performance is the correlation the researchers compute across respondents, and it is the useful part: adoption relates positively to throughput and product performance and negatively to delivery stability. The mechanism the authors give is specific — without strong automated testing, mature version control and fast feedback loops, higher change volume produces instability, and teams in loosely coupled architectures see gains while tightly coupled ones see little or none. Google Cloud publishes this and sells the tools it asks about.", "url": "https://flyvolo.ai/en/changes/ev-20251022-devops-engineer-1" }, { "id": "ev-20251104-registered-nurse-1", "occupation_slug": "registered-nurse", "task_ids": "documentation", "title": "Mercy said Microsoft Dragon Copilot's ambient nursing tool is in use on inpatient units at hospitals in St. Louis, Springfield and Fort Smith, turning narrated care into Epic flowsheet entries", "stage": "deployment", "occurred_on": "2025-11-04", "verified_on": "2026-09-10", "source_name": "Mercy — newsroom", "source_url": "https://www.mercy.net/newsroom/2025-11-02/mercy-advances-groundbreaking-ai-tool-for-nursing-through-collab/", "source_tier": "primary", "scope": "One US health system, inpatient units, early rollout; Mercy reports 8–24 minutes saved per shift for high-use nurses and a 29% cut in incremental overtime. Mercy co-developed the tool with Microsoft, so the figures come from an interested party. Nurses review and edit before filing; documentation only, not clinical decisions.", "url": "https://flyvolo.ai/en/changes/ev-20251104-registered-nurse-1" }, { "id": "ev-20251106-chef-1", "occupation_slug": "chef", "task_ids": "fixed-menu-high-volume; supervising-kitchen-automation", "title": "Sweetgreen agreed to sell Spyce, the unit behind its Infinite Kitchen automated makeline, to Wonder for $186.4 million, keeping the technology in its 20+ automated restaurants under a supply deal", "stage": "deployment", "occurred_on": "2025-11-06", "verified_on": "2026-09-10", "source_name": "Sweetgreen — press release (investor relations)", "source_url": "https://s28.q4cdn.com/367108596/files/doc_news/Sweetgreen-Announces-Strategic-Sale-of-Spyce-to-Wonder-2025.pdf", "source_tier": "primary", "scope": "One US fast-casual chain (bowls and salads): assembly-line automation of a fixed menu, with 38 Spyce staff moving to the buyer. Company release; earlier earnings calls cited labour savings, this release gives none.", "url": "https://flyvolo.ai/en/changes/ev-20251106-chef-1" }, { "id": "ev-20251119-assembly-line-worker-2", "occupation_slug": "assembly-line-worker", "task_ids": "machine-tending-and-changeover", "title": "Figure published the results of an 11-month humanoid deployment at BMW Spartanburg: one station, sheet-metal loading, 1,250+ hours of runtime and 90,000+ parts", "stage": "deployment", "occurred_on": "2025-11-19", "verified_on": "2026-09-11", "source_name": "Figure", "source_url": "https://www.figure.ai/news/production-at-bmw", "source_tier": "primary", "scope": "Vendor-published, and deliberately narrow: one use case (picking sheet-metal parts from racks and placing them on a welding fixture) at one plant, on a 10-hour weekday shift. 1,250 hours of runtime over eleven months is a fraction of one person's working year, and the piece is written as a retirement note — Figure 02 has gone back to headquarters and is being replaced by Figure 03. Figure names its own hardware failure point, the forearm. BMW's own headcount is not mentioned anywhere.", "url": "https://flyvolo.ai/en/changes/ev-20251119-assembly-line-worker-2" }, { "id": "ev-20251120-delivery-rider-2", "occupation_slug": "delivery-rider", "task_ids": "the-ride; robot-and-drone-support", "title": "Uber Eats added Starship's sidewalk robots — about 3,000 units across 270+ locations — starting in Leeds and Sheffield", "stage": "deployment", "occurred_on": "2025-11-20", "verified_on": "2026-09-11", "source_name": "TechCrunch", "source_url": "https://techcrunch.com/2025/11/20/uber-eats-will-use-starship-sidewalk-robots-to-deliver-food-in-the-uk", "source_tier": "secondary", "scope": "Two UK cities in December 2025, 'from select merchants', with additional European markets stated for 2026 and the US for 2027 — so most of the announcement is plan. Uber Eats already worked with two other sidewalk robot companies, Serve and Avride, so this is a third partner rather than a first move. The robots travel no more than two miles and typically deliver in under 30 minutes, which is the constraint that decides where they can work at all. The ~3,000 robots and 270+ locations are Starship's own claim and cover its whole business, most of it university campuses, not the Uber partnership.", "url": "https://flyvolo.ai/en/changes/ev-20251120-delivery-rider-2" }, { "id": "ev-20251125-product-manager-2", "occupation_slug": "product-manager", "task_ids": "writing-it-down", "title": "China's internet regulator announced it had penalised a batch of mobile apps over AI content-labelling failures, with measures including ordered rectification and removal from stores", "stage": "constraint", "occurred_on": "2025-11-25", "verified_on": "2026-09-13", "source_name": "国家互联网信息办公室——《网信部门依法集中查处一批存在人工智能生成合成内容标识违法违规问题的移动互联网应用程序》", "source_url": "https://www.cac.gov.cn/2025-11/25/c_1765795550841819.htm", "source_tier": "primary", "scope": "The regulator's own announcement, which matters because the site already holds the rule and this is the rule being applied. The listed failures divide into two groups and both read as product specifications that were not written. For generation services: no visible label on generated content; no visible label carried into exported files; no implicit label in file metadata containing attribute information, the provider's name or code and a content number; and implicit labels placed in non-conforming positions. For distribution services: not verifying implicit labels, not adding a visible notice around published content, not writing propagation information into file metadata, and not giving users a function with which to declare that content is generated. The stated measures were regulatory interviews, orders to rectify within a deadline, and removal or taking offline. Two limits are firm. The announcement names no application and gives no count — a batch is all it says — so nothing here supports a claim about scale. And it concerns services provided in China; the same feature list is not law elsewhere, though the EU's labelling duty runs in the same direction.", "url": "https://flyvolo.ai/en/changes/ev-20251125-product-manager-2" }, { "id": "ev-20251201-graphic-designer-4", "occupation_slug": "graphic-designer", "task_ids": "asset-production", "title": "US graphic design services employment rose for seven months after ChatGPT's release to 59,300, then fell to 50,100 by December 2025 — its level at the depth of the 2020 shutdown", "stage": "labor_impact", "occurred_on": "2025-12-01", "verified_on": "2026-09-12", "source_name": "US Bureau of Labor Statistics — Current Employment Statistics, all employees, graphic design services (NAICS 541430), series CES6054143001", "source_url": "https://data.bls.gov/timeseries/CES6054143001", "source_tier": "primary", "scope": "United States payroll employment at graphic design firms. It is an industry, not an occupation, and the distinction cuts hard here: a designer who moves from an agency to an in-house team at a retailer leaves this series entirely, so a shift in where design happens looks identical to design disappearing. Freelancers, who are a large share of this trade, are not payroll employees and are not counted either. The shape is worth more than the endpoints: the series went up for seven months after November 2022 and turned down from June 2023, so whatever is happening did not begin when the tools were released. The 2023–2025 period was also a broad contraction in marketing spend, which this series cannot separate from anything else. Seasonally adjusted, monthly, in thousands; the most recent months are preliminary.", "url": "https://flyvolo.ai/en/changes/ev-20251201-graphic-designer-4" }, { "id": "ev-20251231-operations-coordinator-2", "occupation_slug": "operations-coordinator", "task_ids": "order-to-fulfilment", "title": "Eurostat: 6.08% of EU enterprises using AI applied it to logistics in 2025, the lowest of the four purposes the dataset publishes", "stage": "deployment", "occurred_on": "2025-12-31", "verified_on": "2026-09-18", "source_name": "Eurostat — Artificial intelligence by size class of enterprise (isoc_eb_ai), EU27, reference year 2025", "source_url": "https://doi.org/10.2908/ISOC_EB_AI", "source_tier": "primary", "scope": "EU27, reference year 2025. The denominator is enterprises using at least one AI technology — themselves 19.95% of enterprises with 10 or more employees — so 6.08% of AI users is about 1.21% of all enterprises, roughly one in eighty. Compared here only against the other three purposes this dataset publishes for 2025: marketing or sales 34.7%, production processes 20.76%, ICT security 19.77%. The codes for business administration, management and human resources carry no values at all for EU27 in 2025, so 'the lowest' is bounded to the four that do have them. Even at enterprises with 250 or more employees logistics reaches 14.95%, against 47.51% for ICT security. 'Logistics' is the survey's category for the enterprise as a whole, not a measurement of one coordinator's work, and the survey cannot see whether a tool replaced a step or sat beside it.", "url": "https://flyvolo.ai/en/changes/ev-20251231-operations-coordinator-2" }, { "id": "ev-20251231-port-worker-1", "occupation_slug": "port-worker", "task_ids": "horizontal-transport; keeping-the-fleet-running", "title": "PSA reported that at its fully automated Tuas terminal traditional prime movers have been replaced by driverless guided vehicles, and that over 1,700 employees moved into new roles since it opened", "stage": "deployment", "occurred_on": "2025-12-31", "verified_on": "2026-09-20", "source_name": "PSA Corporation Limited — Sustainability 2025 @ PSA SG (financial year 2025)", "source_url": "https://www.singaporepsa.com/wp-content/uploads/2026/06/PSA-SG-Sustainability-Report-2025.pdf", "source_tier": "primary", "scope": "One operator, in Singapore, reporting on its own terminals, in an annual sustainability report covering the calendar year 2025. Two clauses carry the linked tasks. On horizontal transport, the report states that at the fully automated Tuas Port traditional prime movers have been replaced by a fleet of driverless automated guided vehicles, guided by a central management network supported by specialised backend systems and dedicated information technology teams. On the work that automation created, it states that new information technology employees undergo a two-week immersion within the control centre to correlate backend system logs with live terminal manoeuvres, and names the training built for this transition, including a work-study diploma in port automation technology run with a technical institute. The mechanism matters more than the fact: this terminal was built around the machines rather than converted, and at the same company's older Pasir Panjang terminal the equivalent vehicles were still in operational trials through 2025 with deployment planned with safety drivers, so the same operator runs both models in one city. What it does not establish, and this is where a reader should be careful: the report gives no quayside headcount, no ratio of control-room staff to machines, and names neither the origin nor the destination roles of the 1,700. Group headcount rose from 9,695 to 10,290 across three years, but that figure spans warehousing, engineering and technology, so it is not evidence about this occupation. Read it beside the number in the same table that moves the other way: collective bargaining agreement coverage fell from 74.3% to 68.3% to 67%, which the report attributes to a growing number of contract staff. The publisher is marketing this terminal as the most automated in the world and the conclusion that automation created good jobs is one the document exists to support.", "url": "https://flyvolo.ai/en/changes/ev-20251231-port-worker-1" }, { "id": "ev-20251231-real-estate-agent-3", "occupation_slug": "real-estate-agent", "task_ids": "finding-the-next-client", "title": "RE/MAX told investors it launched an AI marketing platform for its affiliated agents in 2025, in the filing that attributes a 4.6 percent fall in US and Canada agent count to the housing market", "stage": "deployment", "occurred_on": "2025-12-31", "verified_on": "2026-09-22", "source_name": "RE/MAX Holdings, Inc. — Form 10-K for the year ended 31 December 2025, filed 19 February 2026", "source_url": "https://www.sec.gov/Archives/edgar/data/1581091/000110465926017561/rmax-20251231x10k.htm", "source_tier": "primary", "scope": "The company's own annual report, not a press write-up. The launch is dated only to the year, so it is recorded at the year end the filing covers. What the filing says in its own words: in 2025 it launched Marketing as a Service, a platform for brand affiliates across the US and Canada to market listings, engage clients and drive business growth, described as an AI-enabled system whose contents include automated listing packages, AI generated marketing videos, customizable ad programs and real-time analytics; and its BoldTrail platform now offers an agent-level, email-integrated AI productivity tool called Folio, alongside recruiting and back-office products. The reason this attaches to prospecting rather than to any other task on this page is that those are the things an agent does to stay the name somebody thinks of: campaigns, listing promotion and follow-up. Three things the filing does not give, and the third is the one that matters most. It gives no usage: not how many of the 148,660 agents have touched the platform, not what it produced, not one measure of effect. It is a franchisor rather than an employer — RE/MAX affiliates are independent, the network reports agent counts that franchisees self-report, and the filing says so — which makes this the operator of a network changing what the work is done with, not a boss issuing a tool. And the headcount in the same document must not be read as a consequence of the first paragraph: total agent count rose 1.4 percent to 148,660 while the US and Canada combined fell 4.6 percent to 72,977, with US company-owned regions at 41,998 against 48,401 two years earlier, and the company attributes those declines to challenging housing and mortgage market conditions and broader economic uncertainty. Its own forward-looking risk section separately says that advances in AI and related technology may accelerate the development of tools that diminish the perceived value of full-service real estate agents — a statement about what may happen, filed among the risks, and not a measurement of anything that has.", "url": "https://flyvolo.ai/en/changes/ev-20251231-real-estate-agent-3" }, { "id": "ev-20260113-bank-teller-3", "occupation_slug": "bank-teller", "task_ids": "routine-transactions", "title": "In 2025 over 11,000 Chinese bank branches were approved to close and over 8,400 to open, a net fall of about 2,000 — but rural credit cooperatives and village banks account for more than all of it", "stage": "labor_impact", "occurred_on": "2026-01-13", "verified_on": "2026-09-11", "source_name": "中国证券报·中证网(依据金融监管总局「金融许可证信息」栏目统计)", "source_url": "https://www.cs.com.cn/yh/04/202601/t20260113_6532923.html", "source_tier": "secondary", "scope": "Mainland China, counted from the regulator's own financial-licence register. The net fall is dominated by consolidation rather than digitisation: rural credit cooperatives shed about 2,200 branches, village banks nearly 1,000 and rural cooperative banks nearly 100, driven by provincial rural-bank restructuring and the merger of small institutions. More than 8,400 new branches were still approved in the same year, so the branch network is being redistributed, not simply shrunk.", "url": "https://flyvolo.ai/en/changes/ev-20260113-bank-teller-3" }, { "id": "ev-20260120-translator-3", "occupation_slug": "translator", "task_ids": "bulk-translation", "title": "Netflix told shareholders it is using AI to improve subtitle localization across its catalogue", "stage": "deployment", "occurred_on": "2026-01-20", "verified_on": "2026-09-11", "source_name": "Netflix (Q4 2025 shareholder letter)", "source_url": "https://s22.q4cdn.com/959853165/files/doc_financials/2025/q4/FINAL-Q4-25-Shareholder-Letter.pdf", "source_tier": "primary", "scope": "One line in a shareholder letter, stated in the present tense: \"In content production and promotion, we're using AI to improve subtitle localization, making it easier for our titles to reach more viewers around the world.\" Netflix does not say which languages, what share of titles, which step of the pipeline, or whether any human localiser's work changed. \"Improve\" is the company's word and could mean a tool handed to translators as easily as a draft generated without one. Subtitles only — the letter says nothing about dubbing.", "url": "https://flyvolo.ai/en/changes/ev-20260120-translator-3" }, { "id": "ev-20260122-devops-engineer-3", "occupation_slug": "devops-engineer", "task_ids": "writing-the-config", "title": "Cloudflare's policy automation platform generated a routing policy that leaked BGP prefixes for 25 minutes; the stated fix adds automated policy checks to CI/CD, not human review", "stage": "deployment", "occurred_on": "2026-01-22", "verified_on": "2026-09-18", "source_name": "Cloudflare — Route leak incident on January 22, 2026 (Bryton Herdes, Tom Strickx)", "source_url": "https://blog.cloudflare.com/route-leak-incident-january-22-2026/", "source_tier": "primary", "scope": "One company, one edge router in Miami, 25 minutes (20:25-20:50 UTC). What the automation authored is the point: a change pushed through Cloudflare's policy automation platform removed prefix-list references and left a policy term so permissive that it marked every internal prefix as acceptable to advertise. A human reverted it and paused automation on that router. Recorded as deployment rather than constraint because adoption was not suppressed: the remediation Cloudflare states is adding automatic routing policy evaluation into its CI/CD pipelines — a machine checking the machine, not a human gate. It establishes that routing policy at one large network operator is authored by an automation platform running in production; it does not say what share of that operator's configuration is machine-written, and one incident at one provider is not a measure of the industry.", "url": "https://flyvolo.ai/en/changes/ev-20260122-devops-engineer-3" }, { "id": "ev-20260122-graphic-designer-2", "occupation_slug": "graphic-designer", "task_ids": "asset-production", "title": "WPP merged its production teams from Hogarth, Ogilvy and VML into one ~10,000-person unit running on a single AI workflow platform", "stage": "deployment", "occurred_on": "2026-01-22", "verified_on": "2026-09-11", "source_name": "WPP", "source_url": "https://www.wpp.com/en/news/2026/01/wpp-launches-wpp-production-empowering-clients-to-reimagine-growth-with-world-class-content", "source_tier": "primary", "scope": "One holding group's production arm — advertising content broadly, not graphic design alone — across 40+ cities, effective 23 February 2026. What is on the record is the operating model: every producer moves onto WPP Open's AI-powered workflows. WPP does not say how much of the output is machine-generated, and the release makes no headcount claim in either direction.", "url": "https://flyvolo.ai/en/changes/ev-20260122-graphic-designer-2" }, { "id": "ev-20260122-hr-recruiter-5", "occupation_slug": "hr-recruiter", "task_ids": "sourcing-and-screening; assessment-and-decision", "title": "Korea's AI Framework Act, in force since January 2026, names recruitment first among the judgements it calls high-impact and requires the operator to keep such a system under human supervision", "stage": "constraint", "occurred_on": "2026-01-22", "verified_on": "2026-09-20", "source_name": "국가법령정보센터 — 인공지능 발전과 신뢰 기반 조성 등에 관한 기본법 제2조제4호 사목·제34조·제35조 (시행 2026-01-22)", "source_url": "https://www.law.go.kr/LSW/lsInfoP.do?lsiSeq=268543&efYd=20260122", "source_tier": "primary", "scope": "South Korea, and it binds the AI business operator rather than the recruiter in the room. Article 2(4)(g) defines the high-impact category to include recruitment, loan screening and other judgements or evaluations that materially affect an individual's relationship of rights and obligations — recruitment is the first thing named in that clause, and the loan-screening half of the same sentence is what another page on this site already rests on. Article 34(1) then attaches duties to an operator providing such a system: a risk-management plan, a method for explaining the final output and the main criteria behind it, a user-protection plan, human management and supervision of the high-impact AI, and documents kept so the measures can be checked. The verb there is shall implement. Read the neighbouring article for contrast, because this is the part a summary always flattens: Article 35(1) says the operator shall endeavour to assess the effect on fundamental rights beforehand, which is a best-efforts duty and not the same thing at all, and the flattening always runs toward making the duty sound stronger than it is. Two further limits on how hard this bites. The specifics are left to a presidential decree that is not in this text, and Article 34(2) says the minister may recommend compliance with the measures it publishes. Breaching Article 34(1) also carries no direct penalty: the route runs through investigation and a corrective order, and only ignoring that order reaches a fine. What it does not establish: nothing about how much hiring in Korea is machine-screened, nothing about headcount, and nothing about whether supervision by a person changes any outcome. Human management and supervision is also not defined in the Act itself. One verification note for anyone checking this source: the law portal answers with a normal page while serving none of the statute to a script, because the article text is loaded by JavaScript — the clauses above were read in a browser, not fetched.", "url": "https://flyvolo.ai/en/changes/ev-20260122-hr-recruiter-5" }, { "id": "ev-20260128-cybersecurity-analyst-3", "occupation_slug": "cybersecurity-analyst", "task_ids": "alert-triage", "title": "CISA's statutory AI inventory lists critical-infrastructure network anomaly detection as deployed and its Security Operation Center anomaly detection as inactive", "stage": "deployment", "occurred_on": "2026-01-28", "verified_on": "2026-09-15", "source_name": "DHS AI Use Case Inventory — CISA(依《Advancing American AI Act》公布的法定清单)", "source_url": "https://www.dhs.gov/ai/use-case-inventory/cisa", "source_tier": "primary", "scope": "One agency, and only the unclassified, non-sensitive use cases it is required to publish under the Advancing American AI Act — classified work is outside the inventory by construction, so absence from this list is not absence from the agency. It is CISA's own statement about its own operations, and CISA runs a security operations centre rather than selling tooling to others, which is why this is recorded as an employer deployment rather than vendor material. Two things sit side by side and both belong on this page: network anomaly detection over critical infrastructure and automated detection of personal data inside cybersecurity data are listed under Deployment, while the use case named Security Operation Center (SOC) Network Anomaly Detection is listed under Inactive, next to confidence scoring for threat indicators. The inventory records status, not reasons: it does not say why the SOC use case stopped, whether something replaced it, how many analysts any of this touches, or what it cost. Statuses read from the 2025 annual update published 28 January 2026.", "url": "https://flyvolo.ai/en/changes/ev-20260128-cybersecurity-analyst-3" }, { "id": "ev-20260128-data-engineer-2", "occupation_slug": "data-engineer", "task_ids": "when-the-data-is-wrong", "title": "The US commodities regulator reported in the federal AI inventory that it runs an isolation-forest model daily to flag potentially erroneous data loads", "stage": "deployment", "occurred_on": "2026-01-28", "verified_on": "2026-09-20", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory (entry CFTC-001)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory", "source_tier": "primary", "scope": "One regulator, one dataset, reported under a statutory duty. Entry CFTC-001 in the individually-reported CSV, whose whole text is two sentences: an anomaly detection model \"designed to identify potentially erroneous data loads in TCR data\", using \"an isolation forest model with aggregated data\", which \"automatically runs daily\". The entry carries no operational date, no vendor, no volume and no account of what the model catches or misses, so the date here is the publication of the consolidated inventory (2026-01-28) and nothing about the deployment itself is dated. It establishes that at least one data team has handed the first half of this task, noticing that a load looks wrong, to a scheduled model; it says nothing about the second half, working out what the wrong load means and what to do. The same inventory carries an entry that reads the other way: DOJ-0006, the Bureau of Alcohol, Tobacco, Firearms and Explosives, lists a data platform whose anomaly detection and data-quality features are \"embedded within the tools\" and adds \"while we do not use these features directly\". A listed capability and a used one are different things, and the inventory format lets an agency say which it has.", "url": "https://flyvolo.ai/en/changes/ev-20260128-data-engineer-2" }, { "id": "ev-20260128-machine-learning-engineer-3", "occupation_slug": "machine-learning-engineer", "task_ids": "deciding-good-enough", "title": "The 2025 US federal AI inventory lists 227 deployed systems their agencies rated high-impact; testing is recorded as complete for 44, in progress for 81 and left blank for 102", "stage": "deployment", "occurred_on": "2026-01-28", "verified_on": "2026-09-20", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory (computed over the individually-reported CSV)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory", "source_tier": "primary", "scope": "Computed here from the individually-reported CSV, the way this site treats a statistics series: the whole file was pulled and the counts were taken locally, so a reader can repeat them. Filter: development_stage is Deployed and is_high_impact is High-impact, which gives 227 of 1,040 deployed rows. The hi_testing_conducted field then reads Yes for 44, In-progress for 81 and is empty for 102. The impact-assessment field reads Yes for 36, In-progress for 90 and is empty for 101. The independent-review field never reads complete as such: 31 name a Chief AI Officer review, 4 an internal independent review, 1 an oversight board, 89 say in progress and 102 are empty. Authority to operate reads Yes for 69, No for 60 and is empty for 98. Federal agencies, not this site, chose the high-impact label and filled these fields, under EO 13960, the Advancing American AI Act and OMB Memorandum M-25-21. What the counts establish is what the agencies wrote down about their own gate before deployment; an empty field is an empty field, not a No, and a Yes is a self-report that no one here has checked. The date is the day OMB published the consolidation (repository created 2026-01-28); the underlying deployments span many years and the file does not give one date for the set.", "url": "https://flyvolo.ai/en/changes/ev-20260128-machine-learning-engineer-3" }, { "id": "ev-20260222-bank-teller-2", "occupation_slug": "bank-teller", "task_ids": "helping-with-complexity; sales-and-referral", "title": "US bank branches showed a net decline of only 400 in 2025 — about 1,400 closures against more than 1,000 openings — the fourth consecutive year with fewer closures than the year before", "stage": "constraint", "occurred_on": "2026-02-22", "verified_on": "2026-09-11", "source_name": "The Financial Brand / Bancography", "source_url": "https://thefinancialbrand.com/news/customer-experience-banking/the-branch-is-dead-is-dead-have-we-reached-national-branch-equilibrium-195842", "source_tier": "secondary", "scope": "United States only — the UK and much of Europe are still closing branches at pace, which is why this record sits alongside the Lloyds closures rather than replacing them. The underlying counts are FDIC and NCUA statistics; the framing is not. The author is president of Bancography, a branch-network planning firm, so the interested party here is arguing that branches matter. Branch counts are also not teller counts: a branch can reopen with half the staff it had.", "url": "https://flyvolo.ai/en/changes/ev-20260222-bank-teller-2" }, { "id": "ev-20260224-retail-cashier-2", "occupation_slug": "retail-cashier", "task_ids": "checkout; customer-help", "title": "Convenience chain Huck's put an AI-native point-of-sale system live in one store and began rolling it out to all 135, with on-screen guidance at the register cutting training time", "stage": "pilot", "occurred_on": "2026-02-24", "verified_on": "2026-09-11", "source_name": "C-Store Dive", "source_url": "https://www.cstoredive.com/news/hucks-rolls-out-ai-native-pos-system/812884", "source_tier": "secondary", "scope": "On the reported date this was live in exactly one store — the flagship in Carmi, Illinois — with a chainwide rollout to 135 locations planned for completion by the end of 2026. US convenience retail, one chain and one vendor (Tote.ai), with a second chain of 96 stores named as in the pipeline. Nothing here removes a checkout lane or a cashier: what the system changes is what the person at the register needs to know, and how centrally the store's settings are controlled.", "url": "https://flyvolo.ai/en/changes/ev-20260224-retail-cashier-2" }, { "id": "ev-20260226-chef-2", "occupation_slug": "chef", "task_ids": "fixed-menu-high-volume", "title": "Miso Robotics' FY2025 SEC annual report discloses $514,798 of Flippy net revenue, lists all three CaliBurger sites as uninstalled, and states substantial doubt about it continuing as a going concern", "stage": "constraint", "occurred_on": "2025-12-31", "verified_on": "2026-09-11", "source_name": "Miso Robotics, Inc. — Form 1-K annual report for FY2025 (SEC EDGAR)", "source_url": "https://www.sec.gov/Archives/edgar/data/1710670/000110465926045031/tm2611517d1_partii.htm", "source_tier": "primary", "scope": "One vendor in one niche — the fry station. The filing lists every Flippy contract by name: all three CaliBurger sites uninstalled between September and December 2024, Nio's terminated in the year it was signed, while White Castle has run since 2022 and two customers signed in 2025 are still running — a contraction, not a stop. Revenue moved the other way, from $384,676 in 2024 to $514,798 in 2025: fewer units in the field, more money per unit, both true at once. The filing also discloses that two customers account for all of it. It says nothing about kitchen automation elsewhere: makelines, woks and prep robots are separate products with separate track records.", "url": "https://flyvolo.ai/en/changes/ev-20260226-chef-2" }, { "id": "ev-20260226-marketing-specialist-2", "occupation_slug": "marketing-specialist", "task_ids": "agent-and-automation-operation; paid-media-operation", "title": "WPP announced Elevate28: the holding company becomes a single company of four units unified by its agentic marketing platform, with £500m of gross annualised cost savings", "stage": "deployment", "occurred_on": "2026-02-26", "verified_on": "2026-09-11", "source_name": "WPP", "source_url": "https://www.wpp.com/en/news/2026/02/strategy-update-and-2025-preliminary-results", "source_tier": "primary", "scope": "One holding group, and most of this is a plan with dates attached to it: stabilise in 2026, build in 2027, accelerate from 2028, with the £500m figure a target rather than a result. What is already happening is the operating model — the four units are named (Media, Creative, Production, Enterprise Solutions), WPP Production launched on 23 February 2026, and WPP Open is the existing platform they are being unified on. WPP attributes its underperformance to organisational complexity, not to AI. No headcount figure appears anywhere in the release.", "url": "https://flyvolo.ai/en/changes/ev-20260226-marketing-specialist-2" }, { "id": "ev-20260301-frontend-developer-2", "occupation_slug": "frontend-developer", "task_ids": "design-to-interface", "title": "Figma reported that in March 2026 it began enforcing AI credit limits on the seats that include Figma Make and Dev Mode, and started selling credit add-ons and pay-as-you-go usage", "stage": "constraint", "occurred_on": "2026-03-01", "verified_on": "2026-09-18", "source_name": "Figma, Inc. — Form 10-Q for the quarter ended 2026-06-30, filed 2026-08-05 (SEC EDGAR)", "source_url": "https://www.sec.gov/Archives/edgar/data/1579878/000162828026053348/fig-20260630.htm", "source_tier": "primary", "scope": "The filing establishes metering, not its effect. Figma states it introduced AI credits across all seats during 2025 and began enforcing the limits in March 2026, selling monthly add-ons or pay-as-you-go usage beyond them; the filing names the month, not a day. It reports no usage figure for Figma Make or Dev Mode, and no measurement of whether any team generated less once the limits bit. Figma sells these seats, so it has an interest in how a pricing change is presented — but the fact recorded here, that generating an interface from a design now costs metered credits rather than being included in the seat, is stated in a filing where overstating carries liability. It says nothing about tools from other vendors, and nothing about in-house pipelines.", "url": "https://flyvolo.ai/en/changes/ev-20260301-frontend-developer-2" }, { "id": "ev-20260302-first-line-manager-3", "occupation_slug": "first-line-manager", "task_ids": "answering-for-the-system; performance-and-feedback", "title": "California's SB 7 would have barred relying solely on an automated system to discipline or dismiss a worker; it passed the legislature, was vetoed, and the veto was sustained on 2 March 2026", "stage": "rule_not_enacted", "occurred_on": "2026-03-02", "verified_on": "2026-09-11", "source_name": "California State Legislature — SB 7 (2025–2026), bill status and Governor's veto message (leginfo.legislature.ca.gov)", "source_url": "https://leginfo.legislature.ca.gov/faces/billStatusClient.xhtml?bill_id=202520260SB7", "source_tier": "primary", "scope": "California. The bill would have required three things: notice to a worker before deploying an automated decision system that makes employment decisions; a prohibition on relying solely on such a system for a disciplinary, termination or deactivation decision; and a right for the worker to request the data the system used. The Governor's veto message gives his reasons: the notification duty was unfocused and would fall on any business using even innocuous tools; the restrictions were overly broad (his example is that barring customer ratings as the primary input removes a tool for rewarding high performers); and the disciplinary and termination scenarios are, in his words, partially covered by forthcoming California Privacy Protection Agency regulations. That last point matters: this record establishes that the specific guardrail SB 7 proposed does not exist, not that California has no protection — separate privacy and anti-discrimination rules are outside its scope.", "url": "https://flyvolo.ai/en/changes/ev-20260302-first-line-manager-3" }, { "id": "ev-20260304-bus-driver-2", "occupation_slug": "bus-driver", "task_ids": "the-drive", "title": "Punggol's autonomous shuttles would open to the public from 1 April 2026 after more than 25,000 km of testing, Singapore's Land Transport Authority announced", "stage": "pilot", "occurred_on": "2026-03-04", "verified_on": "2026-09-23", "source_name": "Land Transport Authority (Singapore) — news release", "source_url": "https://www.lta.gov.sg/content/ltagov/en/newsroom/2026/3/news-releases/autonomous-shuttle-services-in-punggol-open-for-public-rides-fro.html", "source_tier": "primary", "scope": "The authority's release, announced at the Transport Ministry's budget debate: shuttles operated by Grab on two Punggol routes had covered over 25,000 km of testing and carried about 740 people in by-invite rides since 12 January 2026; public rides open from 1 April, free at first, with a flat fare planned from mid 2026. A third route run by ComfortDelGro was still in familiarisation. The same authority runs the bus pilot also recorded on this page, so the two are one programme, not two independent findings; what this one adds is a different operator and vehicle type carrying the public. It says nothing about who is on board or supervising, and it measures rides and feedback, not drivers.", "url": "https://flyvolo.ai/en/changes/ev-20260304-bus-driver-2" }, { "id": "ev-20260305-warehouse-worker-4", "occupation_slug": "warehouse-worker", "task_ids": "goods-movement; picking-and-packing", "title": "JD Logistics reported its goods-to-person automated warehousing in over 20 warehouses at end-2025, against a network of over 1,600 self-operated warehouses", "stage": "deployment", "occurred_on": "2026-03-05", "verified_on": "2026-09-11", "source_name": "JD Logistics — FY2025 annual results announcement, HKEX filing (first-party)", "source_url": "https://www.hkexnews.hk/listedco/listconews/sehk/2026/0305/2026030500918.pdf", "source_tier": "primary", "scope": "China, from the company's own HKEX filing. The network is over 1,600 self-operated warehouses plus over 2,000 third-party cloud warehouses, with aggregate floor area above 34 million square metres. The 20-plus figure counts one specific system — its self-developed goods-to-person solution, which the company says entered nationwide replication in 2025 — so it is not a measure of how automated the whole network is; other sites run other equipment. The same filing reports over 1,000 unmanned vehicles in regular operation for transfers between delivery stations and delivery zones, and a first overseas automated warehouse in the UK. No headcount is given alongside these figures.", "url": "https://flyvolo.ai/en/changes/ev-20260305-warehouse-worker-4" }, { "id": "ev-20260309-school-teacher-1", "occupation_slug": "school-teacher", "task_ids": "orchestrating-learning-tools; lesson-preparation", "title": "Estonia's AI Leap pilot reached all 154 upper-secondary schools (~20,000 students, ~4,900 teachers); teachers got ChatGPT and Gemini plus training from Aug 2025, and over 60% use them weekly", "stage": "pilot", "occurred_on": "2026-03-09", "verified_on": "2026-09-10", "source_name": "AI Leap Foundation (TI-Hüpe) — programme update", "source_url": "https://tihupe.ee/en/estonia-launches-nationwide-ai-leap-education-program-and-introduces-learning-supportive-ai-in-upper-secondary-schools/", "source_tier": "primary", "scope": "One small country, grades 10–11, a three-year pilot; a Tartu/Stanford/OpenAI impact study is still pending. The student-facing tutor app (Estonian-language, built with OpenAI) only opened in late January 2026, and 47% of activated accounts used it weekly.", "url": "https://flyvolo.ai/en/changes/ev-20260309-school-teacher-1" }, { "id": "ev-20260310-paralegal-2", "occupation_slug": "paralegal", "task_ids": "document-review; drafting-from-precedent", "title": "Legal AI vendor Legora said its research, review and drafting platform is used by tens of thousands of legal professionals at more than 800 law firms and in-house teams in 50+ markets", "stage": "deployment", "occurred_on": "2026-03-10", "verified_on": "2026-09-11", "source_name": "Legora", "source_url": "https://legora.com/newsroom/legora-raises-550-million-series-d-to-fuel-us-growth", "source_tier": "primary", "scope": "Vendor-reported, inside a funding announcement — the numbers are the company's own and unaudited, and a firm counts as a customer whether ten people use it or a thousand. \"Used by\" is not \"does the work\": nothing here says what share of a document review runs through it, or that any paralegal's task list changed. It is one vendor; Harvey and several others sell into the same firms, so this understates the total and overstates this product.", "url": "https://flyvolo.ai/en/changes/ev-20260310-paralegal-2" }, { "id": "ev-20260326-social-worker-1", "occupation_slug": "social-worker", "task_ids": "case-notes", "title": "Singapore's social minister said 106 social service agencies were using Scribe, a free AI tool that turns conversations into case notes", "stage": "deployment", "occurred_on": "2026-03-26", "verified_on": "2026-09-24", "source_name": "Ministry of Social and Family Development — Welcome address by Minister Masagos Zulkifli at the Social Work Academia-Practice Symposium 2026", "source_url": "https://www.msf.gov.sg/media-room/article/welcome-address-by-minister-masagos-zulkifli-at-social-work-academia-practice-symposium-(swaps)-2026", "source_tier": "primary", "scope": "The minister's own address as published by the ministry, 26 March 2026. He said that since early 2025 the ministry has provided eligible social service agencies with free access to Scribe, a multilingual transcription and summarisation tool developed by Open Government Products, which converts conversations into structured case notes; that by reducing manual documentation it frees social workers to be fully present when listening to a struggling parent, engaging a child who needs attention, or conducting home visits; that 106 agencies were using it to date; and that the ministry is not replacing the human touch that defines the profession. It establishes a deployment across agencies in one country; the figure counts agencies, not social workers or sessions, and nothing here measures time saved, accuracy or outcomes for clients.", "url": "https://flyvolo.ai/en/changes/ev-20260326-social-worker-1" }, { "id": "ev-20260401-ai-implementation-lead-4", "occupation_slug": "ai-implementation-lead", "task_ids": "choosing-what-to-try-first; getting-people-to-use-it", "title": "Census BTOS supplement: 18% of US firms used AI in a business function, and 57% of those in three functions or fewer", "stage": "deployment", "occurred_on": "2026-04-01", "verified_on": "2026-09-12", "source_name": "U.S. Census Bureau, Center for Economic Studies (CES-WP-26-25)", "source_url": "https://www2.census.gov/library/working-papers/2026/adrm/ces/CES-WP-26-25.pdf", "source_tier": "primary", "scope": "All US firms, nationally representative; reference period Nov 2025 - Jan 2026. 18% firm-weighted, 32% employment-weighted, expected to reach 22% within six months; 50-60% (60-70% employment-weighted) for very large firms in Information, Professional Services and Finance. Among adopters, the most common functions are Sales and Marketing 52%, Strategy and Business Development 45%, IT 41%. This is a count of firms, not of occupations or of people: it says where AI entered a company, never which roles changed. Two further cautions the authors state themselves - the survey questions were revised between supplements, and 23% of firms report workers using AI in tasks against 18% firm-level adoption, so some use is arriving without a company decision at all.", "url": "https://flyvolo.ai/en/changes/ev-20260401-ai-implementation-lead-4" }, { "id": "ev-20260401-business-systems-owner-3", "occupation_slug": "business-systems-owner", "task_ids": "shaping-the-process-in-the-system; deciding-what-the-system-may-decide", "title": "Census BTOS supplement: 16% of AI-using US firms replaced existing software or equipment with AI-integrated solutions", "stage": "deployment", "occurred_on": "2026-04-01", "verified_on": "2026-09-12", "source_name": "U.S. Census Bureau, Center for Economic Studies (CES-WP-26-25)", "source_url": "https://www2.census.gov/library/working-papers/2026/adrm/ces/CES-WP-26-25.pdf", "source_tier": "primary", "scope": "All US firms, nationally representative; reference period Nov 2025 - Jan 2026. The 16% is of AI-using firms, not of all firms. The survey asks the firm what it replaced; it does not identify which system, which vendor, or who administered it, so nothing here says a particular CRM or order system was swapped. Read beside the same survey's labour figures rather than instead of them: among AI-using firms 95% report no AI-related employment change at all, with increases 2.3% and decreases 2.0% (firm-weighted), while the share reporting a large number of replaced tasks rose from 2.5% to 7% since the first supplement - a rise the authors themselves caution may partly reflect revised question wording.", "url": "https://flyvolo.ai/en/changes/ev-20260401-business-systems-owner-3" }, { "id": "ev-20260401-real-estate-agent-2", "occupation_slug": "real-estate-agent", "task_ids": "answering-for-what-you-said", "title": "Japan's Real Estate Brokerage Act requires a licensed transaction specialist to explain the listed material matters and put their own name on the document before a contract concludes", "stage": "constraint", "occurred_on": "2026-04-01", "verified_on": "2026-09-13", "source_name": "e-Gov 法令検索 (Japan, Digital Agency) — 宅地建物取引業法 第三十五条(重要事項の説明等)", "source_url": "https://laws.e-gov.go.jp/law/327AC1000000176", "source_tier": "primary", "scope": "Read on the Digital Agency's own statute portal, in the version in force from 1 April 2026; the Act itself dates from 1952, which is the most useful thing about it — this requirement was not written with any technology in mind and simply happens to bind one now. Article 35(1) requires the broker to have a 宅地建物取引士, a licensed transaction specialist, deliver a written document and explain at least the listed matters before the contract concludes; the list runs to registered rights over the land, planning and building-code restrictions, private-road burdens, the state of water, electricity, gas and drainage, and for an existing building whether a condition survey has been carried out and what it found. Paragraph 4 requires that specialist to present their licence card when giving the explanation, and paragraph 5 requires them to put their name on the document. Two qualifications keep this from being read as more than it is. The document may be provided by electronic means with the counterparty's consent, which replaces the signature and seal — so the duty is a licensed person, not paper and not physical presence. And paragraph 6 disapplies paragraphs 4 and 5 and relaxes the wording where the counterparty is itself a licensed broker: the personal explanation is owed to a consumer, not between professionals. It binds transactions in Japan. It counts nobody and says nothing about how these explanations are prepared.", "url": "https://flyvolo.ai/en/changes/ev-20260401-real-estate-agent-2" }, { "id": "ev-20260401-technical-writer-1", "occupation_slug": "technical-writer", "task_ids": "drafting-the-reference; deciding-what-to-document", "title": "Census BTOS supplement: writing is the leading generative-AI task reported inside US firms, and 65% of firms limit AI use to three tasks or fewer", "stage": "worker_adoption", "occurred_on": "2026-04-01", "verified_on": "2026-09-13", "source_name": "U.S. Census Bureau, Center for Economic Studies (CES-WP-26-25), worker-task findings", "source_url": "https://www2.census.gov/library/working-papers/2026/adrm/ces/CES-WP-26-25.pdf#worker-tasks", "source_tier": "primary", "scope": "Nationally representative, reference period November 2025 to January 2026. This is the worker-task layer of the same supplement rather than its firm-adoption layer: 23% of firms (41% employment-weighted) report workers using AI in work-related tasks, and among the tasks reported, writing, document analysis and information search lead. It is a count of firms reporting a task, not a measure of how much writing is done that way, and it does not distinguish technical documentation from any other writing — nobody asked what was written. Its value here is what it is not: unlike vendor telemetry, this is a statistics agency counting the same channel with no stake in the answer, which makes it the better source for worker adoption where it exists.", "url": "https://flyvolo.ai/en/changes/ev-20260401-technical-writer-1" }, { "id": "ev-20260402-school-teacher-4", "occupation_slug": "school-teacher", "task_ids": "marking-and-feedback; lesson-preparation", "title": "Five Chinese ministries issue an AI-plus-Education action plan requiring AI to be written into the teacher qualification exam and certification, with tiered training to reach every teacher", "stage": "mandate", "occurred_on": "2026-04-02", "verified_on": "2026-09-11", "source_name": "教育部等五部门《关于印发「人工智能+教育」行动计划的通知》教科信〔2026〕1号(教育部政府门户网站)", "source_url": "http://www.moe.gov.cn/srcsite/A16/s3342/202604/t20260410_1433240.html", "source_tier": "primary", "scope": "Mainland China, document number 教科信〔2026〕1号, signed 2 April 2026, with 2030 as the target year. It is a requirement document: it says what must be done and contains nothing about what has been done — no training completion rate, no timetable for revising the teacher qualification exam, no measurement in classrooms. One thing needs saying plainly: the plan itself states that it will continue to refine the 2025 AI literacy guideline already recorded here, so the two come from one policy line rather than two unconnected institutions reaching the same conclusion; do not read them as two pieces of evidence confirming each other. The clauses that bear most directly on teaching are in section seven: lesson preparation is to be supported by automatic generation of multimodal teaching materials, plan optimisation and simulation of the teaching process, and homework handling is to move toward intelligent marking, answering and tutoring.", "url": "https://flyvolo.ai/en/changes/ev-20260402-school-teacher-4" }, { "id": "ev-20260413-administrative-assistant-3", "occupation_slug": "administrative-assistant", "task_ids": "expenses-travel-documents", "title": "Forty-five US federal agencies answered the same 20-item commercial AI checklist: 39 reported an AI tool inside their word processor, 11 reported one for booking travel", "stage": "deployment", "occurred_on": "2026-04-13", "verified_on": "2026-09-22", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory, consolidated commercial off-the-shelf AI use cases (CSV, 45 agencies × 20 questions)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory/blob/3c225ba8438e48306ace7698c8c7feb9486cbc69/Data/2025_consolidated_COTS_AI_use_cases.csv", "source_tier": "primary", "scope": "OMB compiled this file from the inventories each agency publishes under the Advancing American AI Act and OMB Memorandum M-25-21. It is a census rather than a survey: 20 fixed questions, put to every reporting agency, each answered Y or N with the commercial product named and the number of licences given as a band. The file carries 45 agencies and 900 rows; the repository summary states that 46 agencies submitted a consolidated report, so one submission is not in this file. Two of the 20 questions fall inside this task. Using AI-assisted tools in word processors is Y at 39 of 45, joint second of the 20. Finding and booking travel accommodations using AI-powered platforms is Y at 11 of 45, joint sixteenth. Three things this does not establish. The licence band belongs to the product and not to the question: Microsoft 365 Copilot is the product named on many rows at the same agency, so a band of 10,000-50,000 cannot be read as the number of people doing this task with it. Of the 11 agencies reporting AI for travel booking, most name SAP Concur, CWTSatoTravel or an existing government travel system, platforms that predate generative AI, so part of that 11 is agencies pointing at software they already had rather than a tool they went out and bought. And a Y says the agency has such a product, not that any particular assistant uses it, nor how much of the task it does. Read against the rest of the table, the low end is the part worth noticing: scheduling internal meetings and setting reminders is Y at 10 of 45, and logging and analysing time spent on tasks at 9 of 45, eighteenth and nineteenth of the 20.", "url": "https://flyvolo.ai/en/changes/ev-20260413-administrative-assistant-3" }, { "id": "ev-20260413-backend-developer-2", "occupation_slug": "backend-developer", "task_ids": "endpoints-and-plumbing", "title": "Thirty-one of forty-five US federal agencies reported a commercial AI coding tool in the 2025 federal inventory, naming GitHub Copilot, Claude Code, Codex and Poolside", "stage": "deployment", "occurred_on": "2026-04-13", "verified_on": "2026-09-22", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory, consolidated commercial off-the-shelf AI use cases (CSV, 45 agencies × 20 questions)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory/blob/3c225ba8438e48306ace7698c8c7feb9486cbc69/Data/2025_consolidated_COTS_AI_use_cases.csv", "source_tier": "primary", "scope": "The same census: 20 fixed questions put to every reporting agency under a statutory duty, each answered Y or N with the product named and the licences given as a band. Generating code using AI is Y at 31 of 45, seventh of the 20. What makes this row worth more than its count is the product column, because the tools named are ones that did not exist three years ago: GitHub Copilot at Agriculture, Veterans Affairs, the SEC and a dozen more, Claude Code at the Interior, Poolside at Homeland Security and Labor, Codex at the Office of Personnel Management, Amazon Q Developer at Health and Human Services. Compare the travel-booking row in the same file, where most of the Y answers name SAP Concur: on that question agencies were largely pointing at a platform they already had, and on this one they were not. Three limits. A band of 10,000-50,000 at Energy or the Environmental Protection Agency is an enterprise licence that also answers other rows in this file, so it is not a count of people writing code with it. A Y is the employer reporting that it has the tool, not a measurement of how much code the tool writes or how much of that code is kept. And the question says code without distinguishing backend from frontend, or an engineer of ten years from one of ten months; this record attaches here because writing routine service code is what this task is, not because the inventory says so.", "url": "https://flyvolo.ai/en/changes/ev-20260413-backend-developer-2" }, { "id": "ev-20260413-data-engineer-3", "occupation_slug": "data-engineer", "task_ids": "feeding-the-models", "title": "Thirty-five of forty-five US federal agencies reported running a knowledge retrieval system over their own agency information", "stage": "deployment", "occurred_on": "2026-04-13", "verified_on": "2026-09-22", "source_name": "OMB — 2025 Federal Agency AI Use Case Inventory, consolidated commercial off-the-shelf AI use cases (CSV, 45 agencies × 20 questions)", "source_url": "https://github.com/ombegov/2025-Federal-Agency-AI-Use-Case-Inventory/blob/3c225ba8438e48306ace7698c8c7feb9486cbc69/Data/2025_consolidated_COTS_AI_use_cases.csv", "source_tier": "primary", "scope": "The same census, fourth of the 20 questions: searching for agency information using a knowledge retrieval system is Y at 35 of 45. The product column names the systems themselves rather than a category — Microsoft 365 Copilot at most of them, ChatGPT Enterprise at Energy, Claude and Gemini through the GSA USAi platform, Palantir Foundry and Credal at Health and Human Services, Vertex AI and RelativityOne at Homeland Security, AWS Bedrock at Commerce and Transportation. This task is labelled emerging, meaning work that exists because of automation, and what the census establishes is its premise: at 35 of 45 agencies there is a retrieval system pointed at the documents of that agency, so there is a corpus somebody has to keep current and scope. What it does not establish is the work. The file has no column for who maintains what a system reads, no measure of effort, and nothing saying whether that job sits with a data team, a records office or the vendor. It is also one employer and an unusual one: an agency that must publish an inventory of its AI, and must answer for who may see which document, is not a fair stand-in for an employer under neither duty.", "url": "https://flyvolo.ai/en/changes/ev-20260413-data-engineer-3" }, { "id": "ev-20260414-software-tester-2", "occupation_slug": "software-tester", "task_ids": "exploratory-and-adversarial; release-judgement", "title": "A survey of 200 SRE and DevOps leaders reported 43% of AI-generated code changes still need manual debugging in production after passing QA and staging", "stage": "constraint", "occurred_on": "2026-04-14", "verified_on": "2026-09-11", "source_name": "VentureBeat", "source_url": "https://venturebeat.com/technology/43-of-ai-generated-code-changes-need-debugging-in-production-survey-finds", "source_tier": "secondary", "scope": "Self-reported survey of 200 senior engineers at large US, UK and EU enterprises — and the report is published by Lightrun, which sells debugging tools, so the finding and the product point the same way. The Amazon outages it cites (2 and 5 March 2026, traced to AI-assisted changes deployed without approval, followed by a 90-day code safety reset across 335 systems) are independently reported events; the percentages are not. Enterprise software only.", "url": "https://flyvolo.ai/en/changes/ev-20260414-software-tester-2" }, { "id": "ev-20260428-translator-4", "occupation_slug": "translator", "task_ids": "bulk-translation; post-editing", "title": "China's translators association reported a 2025 workforce of 6.867 million and pay rising at over 60% of firms, while industry output edged down to 70.12 billion yuan", "stage": "labor_impact", "occurred_on": "2026-04-28", "verified_on": "2026-09-11", "source_name": "中国翻译协会《2026中国翻译行业发展报告》发布稿(一手)", "source_url": "https://www.tac-online.org.cn/2026-04/28/content_43414810.html", "source_tier": "primary", "scope": "Mainland China, the association's own industry census. Against its 2025 report (2024 data: 70.8 billion yuan output, 6.808 million workers), the workforce rose about 0.9% while output fell about 1%. The headline workforce number is roughly six times the 1.135 million counted as full-time translators, so most of it is not full-time work. The same report puts 2,183 firms with AI translation as their main business and calls human-machine collaboration the industry's basic consensus; the 2024 report had already said over 90% of firms were deploying large models.", "url": "https://flyvolo.ai/en/changes/ev-20260428-translator-4" }, { "id": "ev-20260430-business-systems-owner-2", "occupation_slug": "business-systems-owner", "task_ids": "deciding-what-the-system-may-decide", "title": "China's cyberspace regulator reported 868 filed generative AI services and 530 registered applications as of 30 April 2026, and requires every live application to display which filed service it uses", "stage": "constraint", "occurred_on": "2026-04-30", "verified_on": "2026-09-11", "source_name": "国家互联网信息办公室——《关于发布生成式人工智能服务已备案信息的公告(2026 年 3 月至 4 月)》", "source_url": "https://www.cac.gov.cn/2026-05/13/c_1780413225190669.htm", "source_tier": "primary", "scope": "Mainland China. Under the interim measures for generative AI services, a service is filed with the national regulator and an application that calls a filed model through an API is registered with the provincial one; a live application must show, prominently or on its product detail page, the model name and the filing or listing number. For whoever owns a business system, that makes switching on a generative feature a step with a named, checkable precondition rather than a configuration change. The counts are of filings, not of usage: they say how many services cleared the gate, not how many companies deployed them, how well they work, or whether anyone checked the disclosure. It says nothing about systems outside China.", "url": "https://flyvolo.ai/en/changes/ev-20260430-business-systems-owner-2" }, { "id": "ev-20260501-procurement-specialist-3", "occupation_slug": "procurement-specialist", "task_ids": "finding-and-comparing; the-order-and-the-paperwork", "title": "Anthropic's Economic Index for May 2026 records supplier research and purchase-order handling as delegation-shaped, while negotiation and price analysis sit below the median of all occupational tasks", "stage": "worker_adoption", "occurred_on": "2026-05-01", "verified_on": "2026-09-20", "source_name": "Anthropic Economic Index — Cadences (June 2026 release), dataset for the May 2026 window", "source_url": "https://huggingface.co/datasets/Anthropic/EconomicIndex/tree/main/release_2026_06_26", "source_tier": "primary", "scope": "Conversations on one vendor's consumer and professional product, classified by that vendor to occupational task statements, worldwide, for the month beginning 2026-05-01. The figures below are not printed in the report; they were computed here from the release's own dataset file at release_2026_06_26, filtered to the global rows, and anyone can recompute them. Automation share means the share of conversations whose shape was directive or a feedback loop rather than learning or iteration; it describes how a person used the tool, not whether anyone lost work. For finding and comparing suppliers, the task statement about researching and evaluating suppliers on price, quality, service and reliability reads 57.71% automation on traffic that is 51.22% work. For the order and the paperwork, preparing and processing requisitions and purchase orders reads 67.42% automation on traffic that is 88.30% work, which is the cleanest row in this occupation. The reason to trust the second more than the first is the contamination, and it is the most important thing in this record: the highest automation share here belongs to reviewing catalogues and directories to locate goods, at 80.71%, on traffic that is only 20.52% work, because people researching their own purchases land on the same task statement as buyers. The pattern runs the other way at the bottom. The three statements that are almost entirely work all sit below the median: writing product specifications 45.90% on 90.16% work, negotiating and administering supplier contracts 38.30% on 89.36% work, and analysing price proposals 31.73% on 93.33% work, against a median of 50.00% across 3,650 published task nodes. Read honestly, this says that among the purchasing tasks people actually bring to this tool as work, delegation sits at or below the middle of all occupations, and the tasks that look most automated are the ones consumers are asking about. Stability: the April window preserves the same ordering for seven of these eight statements, with negotiation and price analysis swapping the bottom two places. What it does not establish: nobody in this data said they were a buyer, since the task assignment is made by a classifier; the shares of total usage for these nodes are 0.00% to 0.05%, so the underlying counts are small and are not published; and a task statement absent from the file was suppressed rather than measured at zero. The publisher operates the tool and sells it, so only the party with a commercial interest can see this data at all, which is why a single record of this kind is never enough on this site.", "url": "https://flyvolo.ai/en/changes/ev-20260501-procurement-specialist-3" }, { "id": "ev-20260508-registered-nurse-2", "occupation_slug": "registered-nurse", "task_ids": "documentation", "title": "Abridge made its ambient documentation tool available to nurses across all its health-system clients, more than 250, with several named systems already using it", "stage": "deployment", "occurred_on": "2026-05-08", "verified_on": "2026-09-10", "source_name": "Healthcare IT News", "source_url": "https://www.healthcareitnews.com/news/abridge-releases-ambient-ai-tech-nurses", "source_tier": "secondary", "scope": "US health systems, and the nursing documentation task only. Availability across 250+ clients is not adoption by them — the article names a handful already using it. Says nothing about hands-on care, which is most of a shift.", "url": "https://flyvolo.ai/en/changes/ev-20260508-registered-nurse-2" }, { "id": "ev-20260511-pharmacist-2", "occupation_slug": "pharmacist", "task_ids": "clinical-services; patient-counselling", "title": "Ontario expanded pharmacists' scope from July 2026 to administer six more publicly funded vaccines and to assess and prescribe for nine more common ailments, taking the total to 33", "stage": "constraint", "occurred_on": "2026-05-11", "verified_on": "2026-09-11", "source_name": "Government of Ontario", "source_url": "https://news.ontario.ca/en/release/1007431/ontario-expanding-scope-of-practice-for-pharmacists-and-other-health-professionals", "source_tier": "primary", "scope": "One Canadian province, and a government announcing its own policy — the motive is to relieve pressure on primary care, not to comment on automation. The ailments added are minor ones (dandruff, head lice, warts, mild headache), and prescribing for a minor ailment is not general practice. The province reports 2.4 million assessments under the first 19 ailments and participation by over 99% of Ontario pharmacies, which are its own figures. Recorded here as a constraint because it limits what the dispensing-automation evidence implies, not because it restricts any technology.", "url": "https://flyvolo.ai/en/changes/ev-20260511-pharmacist-2" }, { "id": "ev-20260514-first-line-manager-1", "occupation_slug": "first-line-manager", "task_ids": "answering-for-the-system; performance-and-feedback", "title": "Colorado's SB26-189 gives a person the right to meaningful human review after an automated consequential decision goes against them, with a plain-language explanation due within 30 days", "stage": "constraint", "occurred_on": "2026-05-14", "verified_on": "2026-09-11", "source_name": "Colorado General Assembly, SB26-189 Automated Decision-Making Technology — session law chapter 131, signed 2026-05-14", "source_url": "https://leg.colorado.gov/bills/sb26-189", "source_tier": "primary", "scope": "Colorado only. The statutory definition of a consequential decision names employment, so an adverse decision an employer makes with an automated system falls inside this law. One thing this record does not establish needs saying plainly: the law puts the explanation and the review on the deployer, and says nothing about who inside an employer discharges it — it never mentions supervisors. Attaching it to this occupation's tasks is our inference about where the duty lands in practice: on the layer standing closest to the affected person. The law also does not require the review to change the outcome, only that it be meaningful. Duties phase in, with the developer documentation requirement starting 1 January 2027 and the attorney general's rules on post-adverse-outcome disclosure due by the same date.", "url": "https://flyvolo.ai/en/changes/ev-20260514-first-line-manager-1" }, { "id": "ev-20260514-hr-recruiter-2", "occupation_slug": "hr-recruiter", "task_ids": "assessment-and-decision", "title": "Colorado enacts SB26-189, repealing and re-enacting its 2024 AI Act: after an automated system makes a consequential decision with an adverse outcome, the person may demand meaningful human review", "stage": "constraint", "occurred_on": "2026-05-14", "verified_on": "2026-09-11", "source_name": "Colorado General Assembly, SB26-189 Automated Decision-Making Technology — session law chapter 131, signed 2026-05-14", "source_url": "https://leg.colorado.gov/bills/sb26-189", "source_tier": "primary", "scope": "Colorado only. The statutory definition of a consequential decision names employment explicitly (alongside education, housing, financial and lending services, insurance, health care, and essential government services and benefits). Duties phase in: from 1 January 2027 a developer must give deployers technical documentation covering intended uses, training-data categories, known limitations and instructions for appropriate use and human review; a deployer must give a plain-language description within 30 days of an automated consequential decision that produced an adverse outcome. Enforcement runs through the Colorado Consumer Protection Act, a violation counts as a deceptive trade practice, and before 2030 the attorney general must offer a cure period first. The 2024 act it replaces never took effect. This record makes no claim about exemption thresholds by headcount or revenue — the assembly page does not state them and we did not verify them.", "url": "https://flyvolo.ai/en/changes/ev-20260514-hr-recruiter-2" }, { "id": "ev-20260515-product-designer-2", "occupation_slug": "product-designer", "task_ids": "cross-functional-alignment", "title": "Figma reported paid customers up 54% year on year to about 690,000 and new Pro team conversions up more than 150%, attributing seat expansion to adoption of its AI products", "stage": "constraint", "occurred_on": "2026-05-15", "verified_on": "2026-09-11", "source_name": "Figma, Inc. — Q1 2026 results release, Exhibit 99.1 to Form 8-K filed 2026-05-14 (SEC EDGAR)", "source_url": "https://www.sec.gov/Archives/edgar/data/1579878/000162828026035087/q126pressrelease.htm", "source_tier": "primary", "scope": "One vendor's quarter, reported by the vendor. Seats are not designers: a seat count rising 54% while AI tools spread is at least as consistent with design work moving to people who were never design hires as it is with more designers being employed. Figma also began charging for AI usage on 18 March 2026, so part of the revenue figure is a pricing change rather than growth. The company names its own competitive risk in the same release — Google's Stitch launched free in February, and the CFO said of Anthropic's Claude Design that 'you can't dismiss them'.", "url": "https://flyvolo.ai/en/changes/ev-20260515-product-designer-2" }, { "id": "ev-20260518-ride-hail-driver-3", "occupation_slug": "ride-hail-driver", "task_ids": "the-drive", "title": "Baidu said Apollo Go delivered 3.2 million fully driverless rides in Q1 2026, with weekly rides peaking above 350,000 in March and total rides up over 120% year over year", "stage": "deployment", "occurred_on": "2026-05-18", "verified_on": "2026-09-11", "source_name": "Baidu — Q1 2026 results release (first-party)", "source_url": "https://www.prnewswire.com/news-releases/baidu-announces-first-quarter-2026-results-302774476.html", "source_tier": "primary", "scope": "Baidu's own quarterly disclosure. Cumulative public rides passed 22 million by April 2026 and the service was in 27 cities by May 2026, across mainland China and overseas markets including Dubai; the release reports over 330 million autonomous kilometres, of which over 220 million fully driverless. It gives ride counts, not revenue per ride, utilisation or subsidy levels, and says nothing about the number of drivers in the cities it operates in.", "url": "https://flyvolo.ai/en/changes/ev-20260518-ride-hail-driver-3" }, { "id": "ev-20260521-general-practitioner-2", "occupation_slug": "general-practitioner", "task_ids": "holding-the-pen", "title": "Japan's Medical Practitioners Act forbids a physician from treating a patient or issuing a prescription without personally examining them, and makes doing so a criminal offence", "stage": "constraint", "occurred_on": "2026-05-21", "verified_on": "2026-09-13", "source_name": "e-Gov 法令検索 (Japan, Digital Agency) — 医師法 第二十条・第三十三条の三", "source_url": "https://laws.e-gov.go.jp/law/323AC0000000201", "source_tier": "primary", "scope": "Read on the Digital Agency's own statute portal, in the version in force from 21 May 2026; the Act dates from 1948. Article 20 states that a physician shall not provide treatment, or issue a medical certificate or a prescription, without personally examining the patient; shall not issue a birth or stillbirth certificate without personally attending; and shall not issue a postmortem certificate without personally conducting the examination. One exception is written in: a death certificate for a patient under treatment who dies within twenty-four hours of the consultation. Article 33-3 makes a breach of Article 20 punishable by a fine of up to ¥500,000, so this is a criminal prohibition on the physician personally, not a professional guideline. State plainly what this record does not cover: the Act does not define what personally examining requires, and how it applies to remote or video consultation is set by Ministry of Health, Labour and Welfare guidance that was not read for this record — so nothing here should be taken to mean the physician must be in the room. What it does establish is narrower and harder: the prescription is attached to an act by a named licensed person, and the law does not contemplate that act being performed by anything else. It binds practice in Japan and counts nobody.", "url": "https://flyvolo.ai/en/changes/ev-20260521-general-practitioner-2" }, { "id": "ev-20260522-journalist-3", "occupation_slug": "journalist", "task_ids": "drafting-and-structure", "title": "An arbitrator ruled that POLITICO breached its journalists' union contract by publishing AI-written coverage without editorial review, and the company agreed in May 2026 to shut both tools down", "stage": "constraint", "occurred_on": "2026-05-22", "verified_on": "2026-09-22", "source_name": "PEN Guild (POLITICO and E&E News Guild, NewsGuild-CWA) — release of 22 May 2026", "source_url": "https://www.pen-guild.org/news/victory-politico-agrees-to-shut-down-both-ai-tools-at-center-of-landmark-arbitration", "source_tier": "primary", "scope": "The PEN Guild is the union unit at POLITICO and E&E News and was a party to this arbitration, so this is one side of a bilateral agreement, published by the side that won. What is dated and specific in it: grievances filed in August 2024 over two products deployed without the 60-day notice, the good-faith bargaining and the human oversight that the unit first contract requires; an arbitration hearing on 11 July 2025; a ruling on 26 November 2025 that the collective bargaining agreement had been violated; and an agreement announced on 22 May 2026 under which Capitol AI Report-Builder is shut down and the Live Summaries feature will not be revived. Both products are named, and so is what they produced: unedited live coverage of the 2024 Democratic National Convention and the vice-presidential debate, and branded policy reports sold to POLITICO Pro subscribers that included the claim that Roe v. Wade remains law. Three things to hold onto. The arbitration award itself is not published, so the sentence quoted from it — that AI, as used in these instances, cannot yet rival the hallmarks of human output — is as the union quotes it and not as the document reads. No statement from the employer could be found on its own site or in anything else it publishes, so the agreement to shut the tools down is reported by the other party. And the finding on record is a contract breach rather than a judgement about the work: it establishes that at one newsroom the machine was required to stop, not that it could not do the job. The two sit close together here, because the quoted sentence turns on accuracy, but they are not the same finding.", "url": "https://flyvolo.ai/en/changes/ev-20260522-journalist-3" }, { "id": "ev-20260522-police-officer-1", "occupation_slug": "police-officer", "task_ids": "the-patrol; taking-reports", "title": "Singapore's police began drone patrols from eight pods and put an AI report-lodging assistant into its division headquarters and neighbourhood centres", "stage": "deployment", "occurred_on": "2026-05-22", "verified_on": "2026-09-23", "source_name": "Singapore Police Force — Police Life, From Frontlines to Frontiers: Police Workplan Seminar 2026", "source_url": "https://www.police.gov.sg/media-hub/police-life/2026/05/from-frontlines-to-frontiers-police-workplan-seminar-2026", "source_tier": "primary", "scope": "The force's own account of its Workplan Seminar, published 22 May 2026. It states that its unmanned systems optimise manpower deployment by automating routine patrols and enabling remote operations, with mature systems overseen by a single operator managing several at once and a human in the loop for newer ones; that from May 2026 the Home Team SkyGuardian drones began aerial patrols from eight drone pods, with operators at the Police Operations Command Centre providing remote oversight; that the Report Lodging Co-Pilot, an AI-assisted chatbot at self-help kiosks, was first rolled out at all seven Police Land Division headquarters in October 2025 and had expanded to 21 more Neighbourhood Police Centres by April 2026; that patrol robots will expand to more Changi Airport terminals and transport nodes from 2027; and that the Police Coast Guard is trialling an unmanned surface vessel which since March 2026 has replaced one manned patrol twice a week. It establishes deployments at one national force; it does not say how many officers the force employs, how many reports go through the chatbot, or that any of these systems has reduced headcount. The vessel is a trial, not a deployment.", "url": "https://flyvolo.ai/en/changes/ev-20260522-police-officer-1" }, { "id": "ev-20260522-police-officer-2", "occupation_slug": "police-officer", "task_ids": "taking-reports", "title": "Singapore's Second Minister for Home Affairs said the police cannot hire its way out of a shortage and described AI deployed from report lodging onward", "stage": "deployment", "occurred_on": "2026-05-22", "verified_on": "2026-09-23", "source_name": "Ministry of Home Affairs — Police Workplan Seminar 2026, keynote address by Mr Edwin Tong, Minister for Law and Second Minister for Home Affairs", "source_url": "https://www.mha.gov.sg/media-room/newsroom/police-workplan-seminar-2026/", "source_tier": "primary", "scope": "The minister's own speech as published by the Ministry of Home Affairs, 22 May 2026. He said the force cannot grow indefinitely and that, given Singapore's demographics, it is not possible to throw more officers at a problem, so technology is a necessary force enabler and multiplier; that the Report Lodging Co-Pilot has been progressively deployed since October 2025; that the Case Summariser module of the Investigator Co-Pilot is in pilot implementation, with further modules next year and a Case Recommender in development, framed as sharpening officers' judgement rather than replacing it; that the force is progressively rolling out TRACER, video analytics that processes footage submitted by the public, automatically identifies traffic violations and pinpoints their timestamps; and that two of every three scam websites blocked today are flagged by the force's AI tools. The force's own account of the same day describes TRACER as a capability it is exploring, so the two sources differ on how far along it is. It establishes a minister's statement of the manpower constraint and of what is being deployed; it does not measure any system's effect on staffing or casework.", "url": "https://flyvolo.ai/en/changes/ev-20260522-police-officer-2" }, { "id": "ev-20260522-train-driver-2", "occupation_slug": "train-driver", "task_ids": "being-someone-on-the-system", "title": "Glasgow's transport authority reported that Unite members accepted revised terms on 22 May 2026, and recorded that station staff numbers will increase as the Subway moves to unattended train operation", "stage": "labor_impact", "occurred_on": "2026-05-22", "verified_on": "2026-09-20", "source_name": "Strathclyde Partnership for Transport — Updated Station Staff and Driver Terms and Conditions (report dated 17 June 2026)", "source_url": "https://www.spt.co.uk/media/zgubqqou/p260626_agenda8.pdf", "source_tier": "primary", "scope": "One system, in one city, and the document is a public body's own board paper recording an accepted collective agreement rather than a plan. Under Consequences it states that the number of station staff has been reviewed and will increase to support extended operating hours, and that the agreement carries confirmation that no further industrial action will be taken in relation to these terms and conditions or extended operating hours. Negotiations had been running since 2022, alongside an industrial dispute in 2025, and were deliberately timed to conclude with the modernisation so that all implications of the target operating model could be understood by all parties. The cost is stated: up to around 244,000 pounds per annum, made up of a de-consolidated shift allowance rising from 12% to 14%, better sick pay, higher overtime multipliers and a festive payment rising from 75 to 100 pounds a day. One structural detail matters more than the money: the conditions of service instrument covers station staff and drivers together, in two grades, and it is being renewed into the unattended era rather than wound up. What it does not establish: an increase in station staff is not the same as the drivers being retained, the paper gives no headcount for either, and it is a forward statement in a board paper rather than a measurement of what has happened. The same publisher's own reports also show the readiness milestone for unattended operation moving from July 2026 to the first quarter of 2027 between two consecutive updates, so the trigger for all of this has itself slipped.", "url": "https://flyvolo.ai/en/changes/ev-20260522-train-driver-2" }, { "id": "ev-20260526-radiologist-2", "occupation_slug": "radiologist", "task_ids": "reading-the-routine-study", "title": "US diagnostic radiology training posts rose from 997 to 1,083 across five Matches, filling at 98% or above throughout", "stage": "labor_impact", "occurred_on": "2026-05-26", "verified_on": "2026-09-12", "source_name": "National Resident Matching Program — Results and Data: 2026 Main Residency Match", "source_url": "https://www.nrmp.org/wp-content/uploads/2026/05/Main_Match_Results_and_Data-2026.pdf", "source_tier": "primary", "scope": "United States, and it measures entry to training, not employment: these are positions programmes chose to fund and offer, which is a bet on demand several years out rather than a count of work being done today. The figures are the PGY-2 line for Radiology-Diagnostic in Tables 3 and 8A — 997 (2022), 1,006, 1,017, 1,057, 1,083 (2026), filled by all applicants at 99.9%, 100.0%, 99.9%, 98.6% and 98.4% — the main entry point for the specialty; the smaller PGY-1 line rose 132 to 156 over the same period. A specialty can automate part of its work and still train more people, because volume of imaging and the supply of readers move independently; this record bounds the employment story, it does not settle what machines read.", "url": "https://flyvolo.ai/en/changes/ev-20260526-radiologist-2" }, { "id": "ev-20260528-devops-engineer-4", "occupation_slug": "devops-engineer", "task_ids": "the-pager", "title": "Two Google engineers described the agents their own SRE organisation runs in incident response: alert grouping, handoff documents, postmortem drafts, and in some cases autonomous mitigation", "stage": "deployment", "occurred_on": "2026-05-28", "verified_on": "2026-09-22", "source_name": "Google Cloud — Stevan Malesevic and Christopher Heiser, AI in SRE: where and how Google is deploying agentic AI to improve operations (28 May 2026)", "source_url": "https://cloud.google.com/blog/products/devops-sre/how-google-sre-is-using-agentic-ai-to-improve-operations", "source_tier": "primary", "scope": "Google describing its own operations rather than selling into somebody else, written by a Distinguished Software Engineer and a Distinguished Site Reliability Engineer and published on the company blog. What it states in the past or present tense, which is the half worth recording: the SRE team has developed agents that monitor and improve playbooks from their use during incidents and can generate new playbooks from incidents; an alerting agent that groups, pre-processes and enriches alerts before autonomous handlers take them; agents that consolidate the chat spaces, videos and tracking documents used during an incident, create handoff documents between SREs, draft postmortems, and manage internal and external incident communications; agents created to investigate incidents and, in the words of the post, in some cases to autonomously mitigate issues; and AI Insights, a system that reviews past incidents and feeds what it extracts back to those agents. The rest of the post is plans, and is not recorded. Three limits, and the first is the largest. There is not one number in it: no share of incidents touched, no count of agents running, no before and after, and nothing about how many people are on the rota. Second, the stake is direct and is not hidden by the first-party framing — every component named underneath is a Google product (Gemini, the Agent Development Kit, the Gemini Enterprise Agent Platform, MCP on Google API infrastructure), so this post doubles as a reference architecture for things the publisher sells. Third, it is one operator, and an unusual one: an organisation with twenty years of its own reliability discipline is not a fair stand-in for a team of four keeping a pager. What makes it evidence about this task rather than about tooling is where the agents sit: the work around the decision at three in the morning, assembling the context, handing over, telling people what is happening, writing it up afterwards, is the part that moved, and the post says an agentic approach does not necessarily imply removing people from the process for higher-risk services.", "url": "https://flyvolo.ai/en/changes/ev-20260528-devops-engineer-4" }, { "id": "ev-20260528-lawyer-2", "occupation_slug": "lawyer", "task_ids": "legal-research", "title": "Florida's Supreme Court amends Rule 2.515(d)(2): signing a filing represents that the legal authorities cited exist and are accurate, with sanctions available, effective June 15, 2026", "stage": "constraint", "occurred_on": "2026-05-28", "verified_on": "2026-09-11", "source_name": "Supreme Court of Florida, No. SC2026-0673 — In re: Amendments to Florida Rule of General Practice and Judicial Administration 2.515 (2026-05-28)", "source_url": "https://flcourts-media.flcourts.gov/content/download/2489374/opinion/Opinion_SC2026-0673.pdf", "source_tier": "primary", "scope": "Florida courts only, and three things are easy to misread. First, the rule text never mentions artificial intelligence — the duty is source-neutral: however a filing was produced, the signer answers for whether its citations are real. AI appears only in the Court Commentary. Second, it binds self-represented parties as well as attorneys. Third, the commentary states the amendments were adopted principally to create a statewide, uniform replacement for varied circuit court administrative orders imposing disclosure and certification requirements about the use of artificial intelligence in filings — so it adds one uniform accuracy duty while displacing the AI-specific disclosure orders, which is not simply a tightening. Sanctions may include reprimand, contempt, striking the document, dismissal, costs and attorneys' fees.", "url": "https://flyvolo.ai/en/changes/ev-20260528-lawyer-2" }, { "id": "ev-20260528-sales-account-manager-2", "occupation_slug": "sales-account-manager", "task_ids": "negotiation-and-trust", "title": "Salesforce's CEO said its roughly 15,000 engineers had been flat for two years because of AI, while headcount grew mostly in sales", "stage": "labor_impact", "occurred_on": "2026-05-28", "verified_on": "2026-09-13", "source_name": "Salesforce, Inc. — Q1 FY27 Earnings Conference Call transcript (投资者关系自有发布, 2026-05-27)", "source_url": "https://investor.salesforce.com/files/doc_financials/2027/q1/Salesforce-Q1-FY27-Earnings-Transcript.pdf", "source_tier": "primary", "scope": "The company's own transcript, published on its investor relations site. Read the two halves together, because the second is the one this page is about and the one a headline drops. On engineering: the chief executive states the company has about 83,000 employees, that it has roughly 15,000 engineers and has had roughly 15,000 for about two years, that this is mostly flat because AI has been used to make engineers more efficient and this year coding agents more dramatically so, and that not hiring more engineers is a key part of the margin story. On sales: he states headcount has grown but mostly in one area, sales, and gives the reason — agents can qualify and can provide service, but in sales the company still scales. That is an operator of agent software saying, on an earnings call, which part of the sales job its own agents do not do. Limits: this is one company describing itself to investors, with the incentive that implies; the engineering figure is given conversationally and rounded, not as a disclosed statistic; and no number is given for how much sales headcount grew. It says nothing about sales roles at any other company.", "url": "https://flyvolo.ai/en/changes/ev-20260528-sales-account-manager-2" }, { "id": "ev-20260604-warehouse-worker-2", "occupation_slug": "warehouse-worker", "task_ids": "picking-and-packing", "title": "Amazon's touch-sensing robot Vulcan moved from Spokane to handling more complex picking at its Hamburg fulfilment centre, alongside a €10bn European investment and 25,000 planned fulfilment hires", "stage": "deployment", "occurred_on": "2026-06-04", "verified_on": "2026-09-11", "source_name": "Amazon", "source_url": "https://www.aboutamazon.com/news/operations/amazon-proteus-robot-europe-investment-employee-support", "source_tier": "primary", "scope": "Amazon only, and only two named sites for Vulcan (Spokane, Hamburg) — the company does not say what share of picks it handles at either. Most of the same announcement is plan, not deployment: the natural-language Proteus is 'currently being piloted in Amazon's labs' with European deployment planned for the first half of 2027, and STARK was piloted in Barcelona with 15 European sites planned by 2027. The original Proteus, which moves carts rather than picks, is stated as deployed at 25 US fulfilment centres. The 25,000 hires are a plan with no date.", "url": "https://flyvolo.ai/en/changes/ev-20260604-warehouse-worker-2" }, { "id": "ev-20260609-truck-driver-2", "occupation_slug": "truck-driver", "task_ids": "highway-driving", "title": "PepsiCo signed a multi-year agreement with Gatik to run autonomous freight in its North America supply chain, already operating across Texas, Arizona and Arkansas", "stage": "deployment", "occurred_on": "2026-06-09", "verified_on": "2026-09-11", "source_name": "PepsiCo — press release, 2026-06-08: multi-year agreement with Gatik for autonomous freight in North America", "source_url": "https://www.pepsico.com/en/newsroom/press-releases/2026/pepsico-and-gatik-announce-multi-year-agreement-to-deploy-autonomous-freight-in-north-america", "source_tier": "primary", "scope": "US middle-mile freight between distribution centres, warehouses and stores in three states — not long-haul, not unmapped routes, not the last mile. PepsiCo's release says 'hundreds of pickup and drop-off locations' and does not give a count; it also does not say whether a safety observer rides in the cab on these routes. Gatik's own release of 27 January 2026 states that its driverless trucks 'operate without a driver or safety observer', but names Fortune 50 retailers rather than PepsiCo — so driverless operation for PepsiCo specifically is not established here. Both documents are the companies' own announcements of their own deal.", "url": "https://flyvolo.ai/en/changes/ev-20260609-truck-driver-2" }, { "id": "ev-20260610-airline-pilot-2", "occupation_slug": "airline-pilot", "task_ids": "managing-the-automation", "title": "Airbus describes its Vision Landing, DragonFly and Optimate cockpit-automation research as still in the research phase and \"far from commercial certification\"", "stage": "capability", "occurred_on": "2026-06-10", "verified_on": "2026-09-23", "source_name": "Airbus — Computer vision, automated landing and embedded AI for tomorrow's cockpits (newsroom story, 10 June 2026)", "source_url": "https://www.airbus.com/en/newsroom/stories/2026-06-computer-vision-automated-landing-and-embedded-ai-for-tomorrows-cockpits", "source_tier": "primary", "scope": "Airbus's own newsroom story, published 10 June 2026, describing four of its cockpit-automation research programmes: ATTOL (2018, vision-based airport navigation independent of ground infrastructure such as ILS), DragonFly (launched November 2020, automated emergency operations and pilot assistance), Auto'Mate (camera- and LiDAR-based autonomy research for air-to-air refuelling) and Optimate (2023-present, a three-year project on an A350 flight test airframe set to culminate in a complete automated gate-to-gate mission profile). Airbus states plainly that \"the technology is still in the research phase and far from commercial certification,\" and frames every capability named here as relieving pilots of \"repetitive tactical tasks so they can focus fully on the strategic and safety-critical aspects of the flight,\" not removing them from the loop. This is Airbus describing its own research programme, not an airline's operational experience or a regulator's assessment: Airbus builds the aircraft this occupation flies, and a more capable cockpit automation story is commercially useful to Airbus whether or not any airline ever puts it into production. That stake is why this record is capped at one per occupation and marked as vendor material rather than treated as an independent signal. Nothing here is an airline putting any of this into service, and nothing here changes what FCL.060 still requires of a named pilot's own hands.", "url": "https://flyvolo.ai/en/changes/ev-20260610-airline-pilot-2" }, { "id": "ev-20260626-copywriter-3", "occupation_slug": "copywriter", "task_ids": "volume-copy", "title": "Anthropic's usage telemetry: the median agentic session that produces a blog post contains one human prompt, against 13 turns in chat", "stage": "worker_adoption", "occurred_on": "2026-06-26", "verified_on": "2026-09-12", "source_name": "Anthropic Economic Index report: Cadences", "source_url": "https://www.anthropic.com/research/economic-index-june-2026-report", "source_tier": "primary", "scope": "Measured across Claude chat, Cowork, Claude Code and Anthropic's first-party API during the report's sample period (spring 2026, covering the 15 April US tax deadline). It counts conversations, not people and not employers: 81% of conversations producing a blog or article and 80% of those producing marketing content were classified work-related, but nobody asked whether an employer commissioned, paid for or accepted that output. Higher delegated autonomy on the agentic surface appears across 26 of 31 output types — 0.37 points on a 1-5 scale across all conversations — and it survives holding the model constant: among Sonnet-served conversations the gap is still 0.26 points, which is why the authors read it as a property of the product rather than the model. Anthropic operates the tool being measured and sells it, and it is the only party that can see this data at all - both facts belong beside the numbers. Nothing here is occupation-specific: the mapping from conversation to occupation is a classifier's, not an employer's.", "url": "https://flyvolo.ai/en/changes/ev-20260626-copywriter-3" }, { "id": "ev-20260630-architect-2", "occupation_slug": "architect", "task_ids": "documentation-and-detailing", "title": "Honolulu reported permit decisions on its AI plan-check software averaging 32.5 days against 73, and corrections falling from 23.5 to 7.7 per application", "stage": "pilot", "occurred_on": "2026-06-30", "verified_on": "2026-09-11", "source_name": "City and County of Honolulu, Department of Planning and Permitting — Getting started with CivCheck", "source_url": "https://www.honolulu.gov/dpp/getting-started-with-civcheck/", "source_tier": "primary", "scope": "One US city, residential projects only — single- and two-family homes, duplexes, ADUs and Ohana units. The comparison rests on roughly 40 applications since December 2025, self-selected by applicants who chose to use the tool, which is exactly the population most likely to submit cleanly anyway. The figures are the city's own. A local report said use would become mandatory later in 2026; the department's own pages do not state a mandate or a date, so this record does not establish one. The department is explicit that neither the tool nor the Priority Review programme guarantees approval — other agencies still review.", "url": "https://flyvolo.ai/en/changes/ev-20260630-architect-2" }, { "id": "ev-20260630-content-moderator-1", "occupation_slug": "content-moderator", "task_ids": "the-queue", "title": "TikTok reported that automated systems actioned 94.1 percent of the violating content it removed in the European Union, without human review", "stage": "deployment", "occurred_on": "2026-06-30", "verified_on": "2026-09-22", "source_name": "TikTok — seventh DSA transparency report on content moderation in Europe, announced 31 August 2026 (period 1 January to 30 June 2026)", "source_url": "https://newsroom.tiktok.com/digital-services-act-our-seventh-transparency-report-on-content-moderation-in-europe?lang=en-150", "source_tier": "primary", "scope": "The company publishing its own figures because the Digital Services Act requires a report every six months — which is what makes this measurable at all, and why the numbers exist for platforms in the European Union and almost nowhere else. The period is 1 January to 30 June 2026 and the post is dated 31 August 2026. What it states: around 104 million pieces of content removed for violating the company policies, counted across video, live streams, ads, product listings and comments; and that automated systems actioned 94.1 percent of violating content without human review. Three limits, and the first decides how the number may be used. The denominator is enforcement actions, not working hours and not staff: a platform can action almost everything automatically and still employ many people, because the cases a person still sees are the ones the system could not settle, and those are slower per item. The post publishes no moderator headcount beside the rate, so nothing here supports a claim about employment in either direction. And it is one platform in one jurisdiction; the report also notes that the period includes the launch of its shop in eight further member states, so the content mix inside that 104 million is not constant across the half-year either.", "url": "https://flyvolo.ai/en/changes/ev-20260630-content-moderator-1" }, { "id": "ev-20260630-operations-coordinator-1", "occupation_slug": "operations-coordinator", "task_ids": "order-to-fulfilment; supplier-coordination; documents-and-compliance", "title": "C.H. Robinson's Q2 2026 10-Q shows average headcount down from 12,858 to 11,471 while revenue rose 19%, naming automation and AI to cut manual processes as the restructuring initiative", "stage": "labor_impact", "occurred_on": "2026-06-30", "verified_on": "2026-09-11", "source_name": "C.H. Robinson Worldwide, Inc. — Form 10-Q for the quarter ended 2026-06-30 (SEC EDGAR)", "source_url": "https://www.sec.gov/Archives/edgar/data/1043277/000104327726000031/chrw-20260630.htm", "source_tier": "primary", "scope": "One US-listed freight brokerage. What makes it unusual is that the hardest alternative explanation is ruled out inside the same filing: in the same quarter total revenue rose from $4.14bn to $4.93bn (+19.3%) and income from operations from $216m to $256m (+18.4%), so this is not a demand-driven cut. By segment, North American Surface Transportation went from 5,283 to 4,671 and Global Forwarding from 4,436 to 3,699. What it does not establish needs saying: the filing does not break the reduction down by role or task, so which task was automated is our attribution and not the company's statement; the company itself names natural employee turnover alongside automation as a factor in timing; and a brokerage's coordination work is unusually digitisable because its inputs are already electronic, so this does not transfer to operations roles whose inputs are physical.", "url": "https://flyvolo.ai/en/changes/ev-20260630-operations-coordinator-1" }, { "id": "ev-20260701-architect-3", "occupation_slug": "architect", "task_ids": "visualisation", "title": "US architectural services employment was 203,900 in November 2022 and 204,200 in July 2026, never leaving a 201,100–207,500 band in between", "stage": "labor_impact", "occurred_on": "2026-07-01", "verified_on": "2026-09-12", "source_name": "US Bureau of Labor Statistics — Current Employment Statistics, all employees, architectural services (NAICS 541310), series CES6054131001", "source_url": "https://data.bls.gov/timeseries/CES6054131001", "source_tier": "primary", "scope": "United States payroll employment at architectural services firms, which is an industry and not an occupation: it counts everyone those firms employ — drafters, technologists, interior designers, administrators — and it does not count architects working inside engineering firms, developers, contractors or government. The start month is November 2022 because that is when ChatGPT was released, not because anything happened in the industry that month. The figures are seasonally adjusted, monthly, in thousands; July 2026 is preliminary and subject to revision. Forty-five months is long enough to see a shift in hiring and too short to see one in the size of the profession, and flat employment is compatible with the work inside those firms having changed completely.", "url": "https://flyvolo.ai/en/changes/ev-20260701-architect-3" }, { "id": "ev-20260701-auto-mechanic-1", "occupation_slug": "auto-mechanic", "task_ids": "reading-the-fault; doing-the-repair", "title": "US automotive repair employment rose from 992,600 in November 2022 to a peak of 1,042,800 in February 2026 and was 1,034,700 in July", "stage": "labor_impact", "occurred_on": "2026-07-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES8081110001 (automotive repair and maintenance)", "source_url": "https://data.bls.gov/timeseries/CES8081110001", "source_tier": "primary", "scope": "NAICS 8111, automotive repair and maintenance, all employees, seasonally adjusted. It counts payroll employment at repair businesses and cannot see technicians employed by dealerships' parent operations, fleets or self-employed mechanics. The reason to read it carefully on this page is that the occupation's stated threat is a demand change rather than automation: if electric drivetrains reduce the volume of work arriving, it would appear here as a slow flattening rather than a fall, and the series is not yet long enough past the fleet transition to separate that from an ordinary cycle.", "url": "https://flyvolo.ai/en/changes/ev-20260701-auto-mechanic-1" }, { "id": "ev-20260701-civil-engineer-2", "occupation_slug": "civil-engineer", "task_ids": "signing-the-plans; the-independent-check", "title": "Singapore's Building Control Act, as amended from July 2026, requires a qualified person's signed certificate with every plan and lets major works be approved on an accredited checker's certificate", "stage": "constraint", "occurred_on": "2026-07-01", "verified_on": "2026-09-23", "source_name": "Singapore Statutes Online (Attorney-General's Chambers) — Building Control Act 1989, section 5, current version", "source_url": "https://sso.agc.gov.sg/Act/BCA1989", "source_tier": "primary", "scope": "Section 5 of the Act as published on Singapore Statutes Online, read in the current version. An application for approval of building plans must name the appropriate qualified person appointed to prepare them (5(2)(b)(i)) and, from 1 July 2026, include a certificate signed by that person certifying that he or she prepared those plans (5(2)(e)). For major building works it must also carry an accredited checker's certificate that the plans checked do not show any inadequacy in the key structural elements (5(2)(d)). Section 5(5) lets the Commissioner of Building Control approve plans without checking the plans and design calculations, on the basis of the accredited checker's certificate for major works or the qualified person's certificate in other cases, and 5(6) keeps the right to random checks. It establishes who the law makes answerable for a structural design in Singapore; it says nothing about how the design or the check is produced, whether with software or not, and it counts no one.", "url": "https://flyvolo.ai/en/changes/ev-20260701-civil-engineer-2" }, { "id": "ev-20260701-customer-service-representative-4", "occupation_slug": "customer-service-representative", "task_ids": "faq-answering; triage", "title": "US telephone call-centre employment fell from 385,100 in November 2022 to 277,500 in June 2026, a decline of 28% — but it had already fallen 17% in the 41 months before that", "stage": "labor_impact", "occurred_on": "2026-07-01", "verified_on": "2026-09-11", "source_name": "US Bureau of Labor Statistics — Current Employment Statistics, telephone call centres (NAICS 561422), series CES6056142201", "source_url": "https://data.bls.gov/timeseries/CES6056142201", "source_tier": "primary", "scope": "US payroll employment in the telephone call-centre industry, a monthly series designed for comparison over time, unlike the OEWS occupational survey. Industry, not occupation: many customer service representatives work in banks, retailers and airlines rather than in call centres, and are not in this series. The pre-2022 decline runs at about 5.3% a year and the post-2022 decline at about 8.7% a year, so the trend predates generative AI — offshoring, self-service and IVR were already shrinking it — and the question this series answers is whether the rate changed, not whether AI started it.", "url": "https://flyvolo.ai/en/changes/ev-20260701-customer-service-representative-4" }, { "id": "ev-20260701-hr-recruiter-4", "occupation_slug": "hr-recruiter", "task_ids": "sourcing-and-screening", "title": "US employment placement agency staffing peaked at 280,200 in September 2022 and was 245,000 in July 2026, a fall of 12.6%", "stage": "labor_impact", "occurred_on": "2026-07-01", "verified_on": "2026-09-12", "source_name": "US Bureau of Labor Statistics — Current Employment Statistics, all employees, employment placement agencies (NAICS 561311), series CES6056131101", "source_url": "https://data.bls.gov/timeseries/CES6056131101", "source_tier": "primary", "scope": "United States payroll employment at firms whose business is placing people in jobs — agencies and search firms, not the in-house recruiters who are most of this occupation. That makes it a partial view with one advantage: an agency's headcount tracks how much sourcing and screening is being bought as a service, which is the part of the work under the most pressure from tooling. The confound is large and cannot be separated here: this industry moves with the hiring market, and 2023 to 2025 was a sustained slowdown in hiring, so a recruiter shortage and a recruiter surplus both show up in this line. The peak is September 2022, two months before ChatGPT. Seasonally adjusted, monthly, in thousands; recent months preliminary.", "url": "https://flyvolo.ai/en/changes/ev-20260701-hr-recruiter-4" }, { "id": "ev-20260701-management-consultant-1", "occupation_slug": "management-consultant", "task_ids": "the-deck; making-it-happen", "title": "US management consulting employment rose from 1,515,700 in November 2022 to 1,573,800 in July 2026, and the high point of the series is the latest month", "stage": "labor_impact", "occurred_on": "2026-07-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6054161001 (management consulting services)", "source_url": "https://data.bls.gov/timeseries/CES6054161001", "source_tier": "primary", "scope": "NAICS 54161, management consulting services, all employees, seasonally adjusted, pulled whole from the BLS public API and the slope computed here. This is an industry series and not an occupation series: it counts everyone on a consulting firm's payroll including its own finance and IT staff, and it cannot see the very large number of people who do consulting work inside a client company or through a personal service company. It also says nothing about fee levels, hours billed, or the mix between partners and juniors — a firm can grow its headcount while the pyramid beneath each partner shrinks, and this series would not show it.", "url": "https://flyvolo.ai/en/changes/ev-20260701-management-consultant-1" }, { "id": "ev-20260701-truck-driver-3", "occupation_slug": "truck-driver", "task_ids": "highway-driving", "title": "US long-distance truckload trucking employment peaked at 553,600 in August 2022, three months before ChatGPT, and was 501,400 in July 2026", "stage": "labor_impact", "occurred_on": "2026-07-01", "verified_on": "2026-09-12", "source_name": "US Bureau of Labor Statistics — Current Employment Statistics, all employees, general freight trucking, long-distance, truckload (NAICS 484121), series CES4348412101", "source_url": "https://data.bls.gov/timeseries/CES4348412101", "source_tier": "primary", "scope": "United States payroll employment in long-distance truckload carriers — the segment that autonomous trucking targets first, and deliberately narrower than 'trucking', which includes local delivery and specialised haulage that no driverless programme is aimed at. Read the dates before reading the fall: the peak is August 2022, three months before ChatGPT was released and years before any driverless truck carried freight without a person in the cab at scale, and the whole decline tracks the freight downturn that followed the pandemic shipping boom. This series cannot attribute the fall to anything, and the most-supported explanation in it is a demand cycle, not automation. Owner-operators, who are a large part of this segment, are not payroll employees and are not counted. Seasonally adjusted, monthly, in thousands; recent months preliminary.", "url": "https://flyvolo.ai/en/changes/ev-20260701-truck-driver-3" }, { "id": "ev-20260701-waiter-1", "occupation_slug": "waiter", "task_ids": "taking-the-order; running-the-room", "title": "US restaurant employment rose from 10,853,600 in November 2022 to a series peak of 11,226,500 in May 2026, and was 11,137,500 in July", "stage": "labor_impact", "occurred_on": "2026-07-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES7072250001 (restaurants and other eating places)", "source_url": "https://data.bls.gov/timeseries/CES7072250001", "source_tier": "primary", "scope": "NAICS 7225, restaurants and other eating places, all employees, seasonally adjusted. It covers table service and counter service together, which matters because those two have very different exposure — counter ordering has largely moved to a screen in many markets and table service has not — so a flat aggregate can hide two series moving in opposite directions. It counts jobs and not hours, and in an industry where scheduling is the main lever that is a serious limit: covers per server can rise substantially with no change in this line at all.", "url": "https://flyvolo.ai/en/changes/ev-20260701-waiter-1" }, { "id": "ev-20260701-warehouse-worker-3", "occupation_slug": "warehouse-worker", "task_ids": "goods-movement; picking-and-packing", "title": "US warehousing and storage employment grew 54% from June 2019 to November 2022, then flattened at about 1.85 million through June 2026 while robot fleets scaled", "stage": "labor_impact", "occurred_on": "2026-07-01", "verified_on": "2026-09-11", "source_name": "US Bureau of Labor Statistics — Current Employment Statistics, warehousing and storage (NAICS 493), series CES4349300001", "source_url": "https://data.bls.gov/timeseries/CES4349300001", "source_tier": "primary", "scope": "US payroll employment in the warehousing and storage industry, a monthly series designed for comparison over time. Employment went from 1,224,900 in June 2019 to 1,880,800 in November 2022, then to 1,845,800 in June 2026 — a fall of under 2% over three and a half years. The plateau is confounded: post-pandemic e-commerce normalisation would flatten this series with or without robots. It is industry-wide and says nothing about any single operator, and a flat total is consistent with headcount falling at automated sites and rising elsewhere.", "url": "https://flyvolo.ai/en/changes/ev-20260701-warehouse-worker-3" }, { "id": "ev-20260702-assembly-line-worker-4", "occupation_slug": "assembly-line-worker", "task_ids": "robot-programming-and-maintenance", "title": "China's labour ministry put 12 further new occupations out for comment, among them 'embodied-AI robot application technician'", "stage": "labor_impact", "occurred_on": "2026-07-02", "verified_on": "2026-09-11", "source_name": "中国就业网(人力资源和社会保障部)——《12 个新职业向社会公示》,2026-07-10", "source_url": "https://chinajob.mohrss.gov.cn/h5/c/2026-07-10/569764.shtml", "source_tier": "primary", "scope": "Mainland China, and the occupational classification only. It establishes that the authority of record judges the work of putting embodied-AI robots into use to exist as a distinct occupation — not how many people do it, where, or at what wage, and not that any factory has deployed such robots. The same notice adds work-types including 'agent developer' and 'low-altitude logistics operator'. It is a 公示 open for comment to 17 July 2026, following a call for submissions opened in September 2025; at this date the categories were proposed rather than written into the national occupational catalogue. A new category is a signal about the work that grows around robots, and says nothing on its own about the assembly jobs beside them.", "url": "https://flyvolo.ai/en/changes/ev-20260702-assembly-line-worker-4" }, { "id": "ev-20260702-social-worker-2", "occupation_slug": "social-worker", "task_ids": "spotting-need", "title": "Singapore's social ministry set aside S$15 million over three years to incubate and pilot new technology in social services", "stage": "mandate", "occurred_on": "2026-07-02", "verified_on": "2026-09-24", "source_name": "Ministry of Social and Family Development — MSF Strengthens Partnerships to Chart Next Frontier of Social Service Innovation (media release)", "source_url": "https://www.msf.gov.sg/media-room/article/msf-strengthens-partnerships-to-chart-next-frontier-of-social-service-innovation", "source_tier": "primary", "scope": "The ministry's own media release, 2 July 2026. It states that the minister named three focus areas, the first being strengthening prevention by harnessing data more systematically to identify emerging needs and enable timely intervention; that deploying emerging technologies will be guided by human-centred design, professional oversight, transparent and explainable outputs and safe, evidence-based approaches; that the ministry signed memoranda of understanding with NCS and ST Engineering to co-develop and pilot solutions with social service agencies, focused on early-stage experimentation; that these complement existing scale-ready projects, its case management system CaseCentral and Scribe; and that it will set aside 15 million dollars over three years to incubate and pilot emerging technology projects. It funds pilots; it does not show any tool deployed or any family identified by one. The quotations from the two technology companies are vendors speaking about their own contracts.", "url": "https://flyvolo.ai/en/changes/ev-20260702-social-worker-2" }, { "id": "ev-20260710-electrician-2", "occupation_slug": "electrician", "task_ids": "electrification-work", "title": "Applications for commercial electrical apprenticeships rose more than 70% between 2022 and 2024 on the back of the data-centre buildout, while completion runs near 45%", "stage": "labor_impact", "occurred_on": "2026-07-10", "verified_on": "2026-09-11", "source_name": "Quartz", "source_url": "https://qz.com/apprentice-electrician-data-center-timing-gap-071026", "source_tier": "secondary", "scope": "United States, and the numbers come from several different research firms rather than one dataset — the application figure is Validated Insights, the completion rate is Mathematica, the cooling forecast is Goldman Sachs Research. A journeyman licence requires 8,000 hours (non-union) to 10,000 (IBEW), three to five years, and those are licensing rules rather than estimates. This is about the construction phase: a finished data centre needs far fewer electricians than building one does, which is the part the demand figures do not carry.", "url": "https://flyvolo.ai/en/changes/ev-20260710-electrician-2" }, { "id": "ev-20260716-procurement-specialist-2", "occupation_slug": "procurement-specialist", "task_ids": "finding-and-comparing", "title": "Japan's IT promotion agency surveyed 1,799 companies and reported generative AI in use in 31.3% of procurement departments against 71.9% of IT departments", "stage": "worker_adoption", "occurred_on": "2026-07-16", "verified_on": "2026-09-13", "source_name": "IPA (Information-technology Promotion Agency, Japan) — DX動向2026 (DX Trends 2026), full report PDF", "source_url": "https://www.ipa.go.jp/digital/chousa/dx-trend/rcu1hd0000017uk8-att/dx-trend-2026.pdf", "source_tier": "primary", "scope": "A national agency's firm survey, and the reason it is recorded under worker adoption rather than deployment is written into the same report. Asked how generative AI is used inside the organisation, 63.3% of companies answered that individuals use it for their work and 44.8% that individuals or departments are trialling it, while only 11.9% said it is built into a department's work process and 18.6% into a company-wide service. That is the distinction this stage exists for: people are using the tool, and employers have largely not rebuilt anything around it. IPA is an incorporated administrative agency under Japan's METI and sells none of the tools it counts, which is what makes this a better instrument than vendor telemetry. Two limits matter for this page. The figure resolves to a department, not a task — it establishes that this function is among the least-served, not which of its duties are touched; the report's separate question on purposes puts summarising, translating and proofreading documents at 82.5% and information gathering at 77.0% across all respondents. And the survey ran 17 April to 12 June 2026 in Japan, among companies that answered; 41.6% of firms with 100 or fewer employees report no DX activity at all, so the sample leans toward larger organisations.", "url": "https://flyvolo.ai/en/changes/ev-20260716-procurement-specialist-2" }, { "id": "ev-20260716-video-editor-2", "occupation_slug": "video-editor", "task_ids": "generative-and-hybrid-production", "title": "Netflix told shareholders GenAI workflows had been used in roughly 300 of its 2026 titles, with the largest concentration of the work in post-production", "stage": "deployment", "occurred_on": "2026-07-16", "verified_on": "2026-09-11", "source_name": "Netflix (Q2 2026 shareholder letter)", "source_url": "https://s22.q4cdn.com/959853165/files/doc_financials/2026/q2/FINAL-Q2-26-Shareholder-Letter.pdf", "source_tier": "primary", "scope": "One company's own count, in a shareholder letter, with no definition of what counts as 'used' — a single generated establishing shot and a whole sequence both qualify. Netflix names three titles (Glory, Brasil 70: A Saga do Tri, The American Experiment) and three uses: enhanced crowds, historical battle sequences, worldbuilding establishing shots. The letter also makes the opposite point in the same paragraph: in some cases the shots 'would have had to be left out' without the technology, so part of this is work that did not previously exist at that budget.", "url": "https://flyvolo.ai/en/changes/ev-20260716-video-editor-2" }, { "id": "ev-20260720-data-analyst-2", "occupation_slug": "data-analyst", "task_ids": "query-writing; data-trust", "title": "On Beaver, a text-to-SQL benchmark built from real corporate warehouse query logs, a plain LLM scored zero and an agentic setup reached about 10%, against 80-90%+ on the public benchmarks", "stage": "constraint", "occurred_on": "2026-07-20", "verified_on": "2026-09-11", "source_name": "BLOG@CACM (Stonebraker & Chen, MIT)", "source_url": "https://cacm.acm.org/blogcacm/if-you-think-you-can-do-real-world-text-to-sql", "source_tier": "primary", "scope": "Written by the authors of the benchmark, who also have a competing system (Rubicon) to promote — so the interested party here is arguing that the easy benchmarks are wrong. The underlying claim is checkable: Beaver is built from real query logs at MIT's 1,400-table Oracle warehouse and three others, and its leaderboard is public. The four reasons given are structural rather than about model quality — public benchmark data is in the training corpus, real schemas rot into six different columns named 'salary', warehouses carry local idiom, and real queries join two or three tables rather than one.", "url": "https://flyvolo.ai/en/changes/ev-20260720-data-analyst-2" }, { "id": "ev-20260721-compliance-officer-2", "occupation_slug": "compliance-officer", "task_ids": "the-evidence-file", "title": "Korea's AI framework act requires operators above a compute threshold to run a risk-management system and submit their implementation results to the ministry", "stage": "mandate", "occurred_on": "2026-07-21", "verified_on": "2026-09-13", "source_name": "국가법령정보센터 (Korea Ministry of Government Legislation) — 인공지능 발전과 신뢰 기반 조성 등에 관한 기본법 제32조·제35조", "source_url": "https://www.law.go.kr/법령/인공지능발전과신뢰기반조성등에관한기본법/제32조", "source_tier": "primary", "scope": "Read article by article on the Ministry of Government Legislation's portal, in the version in force from 21 July 2026. The finding is the difference in verb strength between two articles, because that difference is what a summary destroys. Article 32 is a hard duty — 하여야 한다: an operator of a system whose cumulative training compute exceeds a threshold set by Presidential Decree shall identify, assess and mitigate risk across the AI lifecycle and shall build a risk-management system that monitors and responds to AI safety incidents, and shall submit the results of doing so to the Minister of Science and ICT, who sets the method by public notice. Article 35 is best-efforts — 노력하여야 한다: an operator providing a high-impact AI product or service shall endeavour to assess in advance its effect on fundamental rights, taking account of vulnerable groups; only where a state body procures such a service must it give preference to one that has been assessed. Article 36 adds that a foreign operator above user or revenue thresholds must appoint a domestic representative, and that a violation by that representative is treated as the appointing operator's own act. Recorded as a mandate rather than a constraint because the act requires a function to exist rather than suppressing adoption. It binds operators in Korea. Nothing here counts a compliance officer, a filing, or an enforcement action; the compute threshold itself is left to a decree this record does not cover.", "url": "https://flyvolo.ai/en/changes/ev-20260721-compliance-officer-2" }, { "id": "ev-20260721-government-service-clerk-3", "occupation_slug": "government-service-clerk", "task_ids": "checking-it-against-the-rule; answering-for-what-the-platform-decided", "title": "Korea's AI Framework Act, in force in its amended form from 21 July 2026, classifies AI used for public-service eligibility decisions as high-impact and requires human management and supervision of it", "stage": "constraint", "occurred_on": "2026-07-21", "verified_on": "2026-09-12", "source_name": "국가법령정보센터 (Korea Ministry of Government Legislation) — 인공지능 발전과 신뢰 기반 조성 등에 관한 기본법, 법률 제21311호", "source_url": "https://www.law.go.kr/%EB%B2%95%EB%A0%B9/%EC%9D%B8%EA%B3%B5%EC%A7%80%EB%8A%A5%EB%B0%9C%EC%A0%84%EA%B3%BC%EC%8B%A0%EB%A2%B0%EA%B8%B0%EB%B0%98%EC%A1%B0%EC%84%B1%EB%93%B1%EC%97%90%EA%B4%80%ED%95%9C%EA%B8%B0%EB%B3%B8%EB%B2%95", "source_tier": "primary", "scope": "South Korea. Article 2(4)(ja) puts decisions by state bodies on eligibility, entitlement and charging for public services inside the definition of high-impact AI; Article 34(1) then requires a provider using such a system to run a risk-management plan, to be able to explain the result and the main criteria behind it, to protect users, to keep documentation — and, item 4, to place the system under human management and supervision. Read the verbs: Article 34 is a duty (shall implement, with the detail left to Presidential Decree), while the fundamental-rights impact assessment in Article 35(1) is only a best-efforts obligation (shall endeavour), and Article 35(2) merely tells state bodies to give priority to assessed products. So the hard requirement is supervision and explainability, not assessment. The Decree fixing the detail is where the real burden will be set and is not in this text. Nothing here says how many counter posts exist or what a clerk does; it fixes what may not be decided by a machine alone.", "url": "https://flyvolo.ai/en/changes/ev-20260721-government-service-clerk-3" }, { "id": "ev-20260721-radiologist-3", "occupation_slug": "radiologist", "task_ids": "answering-for-the-machine-that-read-it", "title": "Korea's AI Framework Act places the use of AI medical devices inside high-impact AI, obliging the user to keep it under human management and supervision and to be able to explain its output", "stage": "constraint", "occurred_on": "2026-07-21", "verified_on": "2026-09-12", "source_name": "국가법령정보센터 (Korea Ministry of Government Legislation) — 인공지능 발전과 신뢰 기반 조성 등에 관한 기본법, 법률 제21311호", "source_url": "https://www.law.go.kr/%EB%B2%95%EB%A0%B9/%EC%9D%B8%EA%B3%B5%EC%A7%80%EB%8A%A5%EB%B0%9C%EC%A0%84%EA%B3%BC%EC%8B%A0%EB%A2%B0%EA%B8%B0%EB%B0%98%EC%A1%B0%EC%84%B1%EB%93%B1%EC%97%90%EA%B4%80%ED%95%9C%EA%B8%B0%EB%B3%B8%EB%B2%95", "source_tier": "primary", "scope": "South Korea. The word that matters in Article 2(4)(ra) is use: the high-impact category covers the development and the use of medical devices and digital medical devices, so a hospital running a cleared AI reading tool is an AI-using operator under this Act and not merely a customer of one. Article 34(1) then attaches duties to that operator — a risk-management plan, an ability to explain the output and the main criteria behind it, documentation, and item 4, human management and supervision. Article 31(1) separately requires telling the user in advance that the service runs on AI. Two limits: the detail is left to a Presidential Decree not contained in this text, and the fundamental-rights impact assessment in Article 35 is only a best-efforts duty. This says nothing about how many radiologists a hospital employs, and nothing about accuracy — it fixes who has to be able to answer for the output.", "url": "https://flyvolo.ai/en/changes/ev-20260721-radiologist-3" }, { "id": "ev-20260721-technical-writer-2", "occupation_slug": "technical-writer", "task_ids": "drafting-the-reference; deciding-what-to-document", "title": "Korea's AI framework act requires operators of high-impact AI to write and retain documents evidencing their safety measures, and to run a plan for explaining the system's criteria and training data", "stage": "mandate", "occurred_on": "2026-07-21", "verified_on": "2026-09-13", "source_name": "국가법령정보센터 (Korea Ministry of Government Legislation) — 인공지능 발전과 신뢰 기반 조성 등에 관한 기본법 제34조·제36조", "source_url": "https://www.law.go.kr/법령/인공지능발전과신뢰기반조성등에관한기본법/제34조", "source_tier": "primary", "scope": "Read article by article on the Ministry of Government Legislation's portal, in the version in force from 21 July 2026. Two sub-paragraphs of Article 34(1) bear on this occupation and they do different things. Sub-paragraph 2 requires the operator to establish and implement a plan for explaining — so far as technically possible — the final result the AI produced, the main criteria used to reach it, and an outline of the training data used to develop and operate it. Sub-paragraph 5 requires the writing and retention of documents by which the content of the safety and reliability measures can be verified. The first governs what a reference has to cover; the second removes the discretion to leave something out. Article 36(1)(3) then makes the point about maintenance rather than authorship: a foreign operator above user or revenue thresholds must appoint a domestic representative whose duties include supporting compliance with Article 34(1), expressly including checking those documents for currency and accuracy. Two limits. The act leaves the detail to Presidential Decree and to a ministerial notice, and Article 34(2) says the minister may recommend compliance with that notice — so the specificity of what must be written is not fixed by this record. And it binds only operators of high-impact AI in Korea; nothing here counts a technical writer, a document or an hour.", "url": "https://flyvolo.ai/en/changes/ev-20260721-technical-writer-2" }, { "id": "ev-20260723-civil-engineer-3", "occupation_slug": "civil-engineer", "task_ids": "the-submission; the-model", "title": "Singapore reported more than 140 projects by 300 firms had submitted through CORENET X, and narrowed its October 2026 mandate to projects of 5,000 m² and above", "stage": "deployment", "occurred_on": "2026-07-23", "verified_on": "2026-09-24", "source_name": "Urban Redevelopment Authority and Building and Construction Authority — Circular URA/PB/2026/08-DCG (APPBCA-2026-12), Updates to CORENET X implementation plan", "source_url": "https://www.ura.gov.sg/guidelines/circulars/dc26-08/", "source_tier": "primary", "scope": "The two agencies' own circular of 23 July 2026, addressed to building owners, developers, architects, engineers, registered surveyors and contractors, read in full. It revises the plan set out in the circular of 10 September 2025. It reports that since submission via CORENET X became mandatory for new projects of 30,000 square metres or more on 1 October 2025, more than 140 projects involving 300 firms have submitted through it, with time savings of up to two months, cost savings from reduced abortive work and fewer resubmissions — figures the agencies report about their own system. Citing feedback that the gains are most pronounced for larger and more complex projects, it makes submission via the CORENET X gateway processes mandatory from 1 October 2026 only for new projects of 5,000 square metres or more; projects below 5,000 square metres may continue on CORENET 2 and use CORENET X voluntarily; ongoing projects of 5,000 square metres or more may continue on CORENET 2 until further details are announced. It establishes real use of the digital route and a narrower mandate than first planned; it does not measure how the work inside firms changed, and the savings are not independently verified.", "url": "https://flyvolo.ai/en/changes/ev-20260723-civil-engineer-3" }, { "id": "ev-20260724-school-teacher-2", "occupation_slug": "school-teacher", "task_ids": "explaining-and-adapting-live; orchestrating-learning-tools", "title": "A New York school district paused an approved 60,000 dollar humanoid teaching robot after state officials, the teachers union and parents objected", "stage": "constraint", "occurred_on": "2026-07-24", "verified_on": "2026-09-10", "source_name": "AP News", "source_url": "https://apnews.com/article/new-york-school-artificial-intelligence-robot-teacher-c2126c704104c4eb68b79738630b07df", "source_tier": "secondary", "scope": "One rural district, one stationary robot bought as a teaching aid for a high-school robotics course — not a system that teaches a class. What stopped it was procurement, student-data agreements and union objection, not any limit on what the device can do.", "url": "https://flyvolo.ai/en/changes/ev-20260724-school-teacher-2" }, { "id": "ev-20260728-graphic-designer-3", "occupation_slug": "graphic-designer", "task_ids": "taste-and-defence", "title": "Equinox removed an AI-generated poster from its 2026 'Question Everything But Yourself' campaign and apologised, after months of criticism that the image was dehumanising", "stage": "constraint", "occurred_on": "2026-07-28", "verified_on": "2026-09-11", "source_name": "Athletech News", "source_url": "https://athletechnews.com/equinox-apologizes-for-ai-ad-campaign-image-asian-woman/", "source_tier": "secondary", "scope": "One brand campaign, run by agency Angry Gods from January 2026, whose stated theme was AI-made false realities; the image ran on posters and social channels for months before removal. Equinox first defended the work publicly, then said it had decided to remove the image. This is one campaign in fitness retail, not a measured industry pattern.", "url": "https://flyvolo.ai/en/changes/ev-20260728-graphic-designer-3" }, { "id": "ev-20260729-retail-cashier-3", "occupation_slug": "retail-cashier", "task_ids": "loss-prevention", "title": "Australia's privacy regulator updated its facial recognition guidance for retail settings on 29 July 2026 to reflect the Administrative Review Tribunal's Bunnings decision of 4 February 2026", "stage": "constraint", "occurred_on": "2026-07-29", "verified_on": "2026-09-20", "source_name": "Office of the Australian Information Commissioner — Facial recognition technology: a guide to assessing the privacy risks (July 2026 edition)", "source_url": "https://www.oaic.gov.au/__data/assets/pdf_file/0025/266335/Facial-recognition-technology-A-guide-to-assessing-the-privacy-risks-July-2026.pdf", "source_tier": "primary", "scope": "Australia only, and only the collection of biometric information under the Privacy Act 1988 (Cth). The document is the regulator's own guidance, not a statute, and it says the Act is technology-neutral and neither bans nor permits facial recognition. What it sets out is the route to lawfulness in a shop: collecting sensitive information must be reasonably necessary and consented to, and an entity that cannot obtain valid consent — including where consent is unreasonable or impracticable — must not use the technology unless one of two narrow exceptions applies; before deploying it, an entity is expected to run a documented risk assessment and to give genuine consideration to whether less privacy-intrusive methods could achieve the same outcome. The passage bearing on this task is the guidance's own case study of the Bunnings decision, which records that identifying known offenders let staff be alerted and remove people before an incident, and that Bunnings also used human intervention to verify matches identified by the system, which the guidance says adequately mitigated the risk of acting on a false positive match. Two counter-signals from the same document: the Tribunal accepted that the deployment was effective and that no less intrusive control could identify repeat offenders in stores of that kind, so this is not a record of a ban; and a separate determination against Kmart Australia over facial recognition used against refund fraud remains under review in the same Tribunal, so that half is unsettled. The duties fall on the retailer, not on the person at the till. The document says nothing about how many shops use the technology, nothing about staffing levels, and nothing about any country other than Australia.", "url": "https://flyvolo.ai/en/changes/ev-20260729-retail-cashier-3" }, { "id": "ev-20260731-copywriter-2", "occupation_slug": "copywriter", "task_ids": "volume-copy", "title": "Writing roles fell from about one in six of UK agency job postings to about one in thirty by July 2026, while overall agency postings rose 24 percent", "stage": "labor_impact", "occurred_on": "2026-07-31", "verified_on": "2026-09-10", "source_name": "The Drum(数据来自 Agency by Agency,基于 The Data City 与 Lightcast)", "source_url": "https://www.thedrum.com/news/as-writing-jobs-go-from-1-6-to-1-30-of-agency-postings-are-copywriters-in-trouble", "source_tier": "secondary", "scope": "Agency job postings in the UK, measured as a share of all agency postings — not employment, and not in-house or freelance copywriting. A share can fall because other roles grew; here total postings rose while writing rose far less.", "url": "https://flyvolo.ai/en/changes/ev-20260731-copywriter-2" }, { "id": "ev-20260801-care-worker-1", "occupation_slug": "care-worker", "task_ids": "the-body-work; noticing-the-change", "title": "US home health care employment rose 21.5% from November 2022 to 1,896,400 in August 2026, the highest point in the series", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6562160001 (home health care services)", "source_url": "https://data.bls.gov/timeseries/CES6562160001", "source_tier": "primary", "scope": "NAICS 6216, home health care services, all employees, seasonally adjusted. This is the fastest-growing health industry series we track and it is growing for demographic reasons that have nothing to do with technology, which is the first thing to hold in mind when reading it. It counts payroll employment at home care agencies only: family members providing unpaid care, workers engaged informally, and staff in residential facilities are all outside it, and in most markets those groups are larger than this one. Growth in headcount also says nothing about pay, hours or turnover, which is where this occupation's real pressure sits.", "url": "https://flyvolo.ai/en/changes/ev-20260801-care-worker-1" }, { "id": "ev-20260801-cleaner-1", "occupation_slug": "cleaner", "task_ids": "large-floors; being-scheduled", "title": "US employment in services to buildings rose from 2,215,800 in November 2022 to 2,307,100 in August 2026, within 0.5% of the series high", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6056170001 (services to buildings and dwellings)", "source_url": "https://data.bls.gov/timeseries/CES6056170001", "source_tier": "primary", "scope": "NAICS 5617, services to buildings and dwellings, all employees, seasonally adjusted. This is the industry that employs contract cleaners, and it also contains landscaping and pest control, so it is wider than this occupation. The more important limit is who it cannot see: cleaners employed directly by a building, a hospital or a school are counted in that employer's industry rather than here, and in several markets that is a large share. A flat or rising series is therefore evidence that contract cleaning has not shrunk, not that cleaning headcount overall has held.", "url": "https://flyvolo.ai/en/changes/ev-20260801-cleaner-1" }, { "id": "ev-20260801-construction-worker-1", "occupation_slug": "construction-worker", "task_ids": "building-in-place; the-sequence", "title": "US construction employment rose from 7,863,000 in November 2022 to 8,359,000 in August 2026, and the high point of the series is the latest month", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES2000000001 (construction)", "source_url": "https://data.bls.gov/timeseries/CES2000000001", "source_tier": "primary", "scope": "Total construction employment, all employees, seasonally adjusted. It is the whole sector, so it includes office staff, engineers and managers as well as site trades, and it cannot separate them. Two limits matter for this page. It counts payroll employment, so self-employed and informally engaged workers — a very large share of site labour in many markets — are invisible in it. And it counts jobs rather than hours or output, so a sector that builds more with prefabricated components would show the manufacturing jobs somewhere else entirely, which is precisely the shift this occupation's page argues is the real mechanism.", "url": "https://flyvolo.ai/en/changes/ev-20260801-construction-worker-1" }, { "id": "ev-20260801-general-practitioner-1", "occupation_slug": "general-practitioner", "task_ids": "writing-the-visit-down; working-out-what-is-wrong", "title": "US employment in physicians' offices rose from 2,852,100 in November 2022 to 3,061,500 in August 2026, its highest point in the series", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6562110001 (offices of physicians)", "source_url": "https://data.bls.gov/timeseries/CES6562110001", "source_tier": "primary", "scope": "NAICS 6211, offices of physicians, all employees, seasonally adjusted. This counts everyone a practice employs — receptionists, medical assistants, nurses, billing staff — so a rise here is not evidence that more doctors are employed, and the composition could shift toward support staff without the series moving differently. It also excludes physicians working in hospitals, which is a large share in many systems, and it says nothing about hours, panel size or how appointment length is changing, which is where this occupation's actual pressure shows up.", "url": "https://flyvolo.ai/en/changes/ev-20260801-general-practitioner-1" }, { "id": "ev-20260801-it-support-specialist-1", "occupation_slug": "it-support-specialist", "task_ids": "the-repeat-ticket; finding-out-what-actually-happened", "title": "US computer systems design employment peaked at 2,483,500 in March 2023 and has fallen about 5% since, to 2,362,700 in August 2026", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6054150001 (computer systems design and related services)", "source_url": "https://data.bls.gov/timeseries/CES6054150001", "source_tier": "primary", "scope": "NAICS 5415, computer systems design and related services, all employees, seasonally adjusted. The turn is March 2023 and the decline is steady rather than sharp. Two limits decide how much weight this can carry for this occupation. The industry contains software development, consulting and managed services as well as support desks, and cannot separate them — so this is not a support-specific measurement. And most people doing IT support are employed inside a bank, a hospital or a factory rather than by an IT services firm, and every one of them is counted in that employer's industry instead. The series sees the outsourced half of this occupation and not the larger in-house half.", "url": "https://flyvolo.ai/en/changes/ev-20260801-it-support-specialist-1" }, { "id": "ev-20260801-lab-technician-1", "occupation_slug": "lab-technician", "task_ids": "running-the-analyser; knowing-the-result-is-wrong", "title": "US medical and diagnostic laboratory employment was 310,200 in August 2026 — below its March 2022 peak but still about 8% above its pre-pandemic level", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6562150001 (medical and diagnostic laboratories)", "source_url": "https://data.bls.gov/timeseries/CES6562150001", "source_tier": "primary", "scope": "NAICS 6215, medical and diagnostic laboratories, all employees, seasonally adjusted. The decline from the March 2022 peak is real and must not be read as automation: the peak is a COVID testing surge (288,100 in December 2019, rising to 322,300 by March 2022) and what has happened since is that surge unwinding, with employment settling above where it started rather than below. Anyone quoting the fall from the peak without the pre-pandemic baseline gets the direction of this series wrong. It also covers standalone laboratories only — technicians working inside a hospital laboratory, which is where most of them are, are counted in hospitals instead.", "url": "https://flyvolo.ai/en/changes/ev-20260801-lab-technician-1" }, { "id": "ev-20260801-loan-officer-1", "occupation_slug": "loan-officer", "task_ids": "scoring-the-application; gathering-the-case", "title": "US credit intermediation employment peaked in March 2021 — twenty months before ChatGPT — and has fallen about 5% since November 2022 to 2,528,300", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES5552200001 (credit intermediation and related activities)", "source_url": "https://data.bls.gov/timeseries/CES5552200001", "source_tier": "primary", "scope": "NAICS 522, credit intermediation and related activities, all employees, seasonally adjusted. Read the peak date before the decline: the high point is March 2021, which is twenty months before ChatGPT was released, so the best-supported explanation for the turn in this series is the end of the mortgage refinancing boom and the interest-rate cycle, not automation. This is an industry series covering everyone a lender employs, including branch and back-office staff, so it cannot isolate lending officers; and it cannot see brokers and originators who are not on a bank payroll.", "url": "https://flyvolo.ai/en/changes/ev-20260801-loan-officer-1" }, { "id": "ev-20260801-marketing-specialist-3", "occupation_slug": "marketing-specialist", "task_ids": "content-and-creative-production", "title": "US advertising and public relations employment peaked at 503,200 in February 2023 and fell every year after, reaching 476,000 in August 2026", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-12", "source_name": "US Bureau of Labor Statistics — Current Employment Statistics, all employees, advertising, public relations and related services (NAICS 5418), series CES6054180001", "source_url": "https://data.bls.gov/timeseries/CES6054180001", "source_tier": "primary", "scope": "United States payroll employment at agencies — advertising, PR, media buying, display and direct mail in one number. Marketers employed by the brands themselves are counted in their employer's industry and not here, so the long shift of marketing in-house, which began well before these tools, shows up in this line as a fall regardless of anything AI did. The peak is February 2023, three months after ChatGPT, and the decline is gradual rather than a break: 503,200 to 476,000 over three and a half years, about 5%. That is a real contraction and a slow one, and the series cannot say how much of it is tools, how much is in-housing, and how much is advertising budgets. Seasonally adjusted, monthly, in thousands; recent months preliminary.", "url": "https://flyvolo.ai/en/changes/ev-20260801-marketing-specialist-3" }, { "id": "ev-20260801-medical-assistant-1", "occupation_slug": "medical-assistant", "task_ids": "vitals-and-intake; the-phone", "title": "US outpatient care centre employment rose 14.2% from November 2022 to 1,194,100 in August 2026, close to its series high", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6562140001 (outpatient care centers)", "source_url": "https://data.bls.gov/timeseries/CES6562140001", "source_tier": "primary", "scope": "NAICS 6214, outpatient care centres, all employees, seasonally adjusted. The industry is growing because care is moving out of hospitals into outpatient settings, which is a structural shift in how care is delivered rather than anything about AI. It counts every employee of those centres — clinicians, assistants, administrators — and cannot separate them, so it is not a count of medical assistants. Its use on this page is narrow: it establishes that the setting this occupation works in is expanding, which is the backdrop against which the paperwork half of the job is being automated.", "url": "https://flyvolo.ai/en/changes/ev-20260801-medical-assistant-1" }, { "id": "ev-20260801-paralegal-3", "occupation_slug": "paralegal", "task_ids": "document-review", "title": "US legal services employment rose from 1,180,400 in November 2022 to 1,245,900 in August 2026, its highest reading in the series", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-12", "source_name": "US Bureau of Labor Statistics — Current Employment Statistics, all employees, legal services (NAICS 5411), series CES6054110001", "source_url": "https://data.bls.gov/timeseries/CES6054110001", "source_tier": "primary", "scope": "United States payroll employment across legal services — lawyers, paralegals, legal assistants and support staff in one number, which is the main limit on reading it here: a firm can grow while shifting its mix away from the paralegal, and this series would not show it. The low point since November 2022 is April 2023 at 1,179,300 and the high is the final month, so the line does not dip at any point in the period during which document review tools became ordinary. In-house legal departments are counted in their employer's industry, not here, so work moving in-house would look like a fall. Seasonally adjusted, monthly, in thousands; the most recent months are preliminary.", "url": "https://flyvolo.ai/en/changes/ev-20260801-paralegal-3" }, { "id": "ev-20260801-physical-therapist-1", "occupation_slug": "physical-therapist", "task_ids": "the-programme; keeping-them-doing-it", "title": "US employment in offices of other health practitioners rose 21.1% from November 2022 to 1,355,200 in August 2026, its series high", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6562130001 (offices of other health practitioners)", "source_url": "https://data.bls.gov/timeseries/CES6562130001", "source_tier": "primary", "scope": "NAICS 6213, offices of other health practitioners, all employees, seasonally adjusted. This industry contains physical, occupational and speech therapy together with chiropractic, optometry and podiatry, so it is considerably wider than this occupation and cannot be read as a count of physiotherapists. It counts practice employees, which includes front-desk and billing staff. And it counts jobs rather than sessions or minutes — the pressure this page describes is the fee schedule shortening what a session contains, and a series of headcounts cannot see that at all.", "url": "https://flyvolo.ai/en/changes/ev-20260801-physical-therapist-1" }, { "id": "ev-20260801-receptionist-1", "occupation_slug": "receptionist", "task_ids": "signing-people-in; the-office-around-it", "title": "US office administrative services employment rose 5.6% from November 2022 to 639,100 in August 2026, the highest point in the series", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6056110001 (office administrative services)", "source_url": "https://data.bls.gov/timeseries/CES6056110001", "source_tier": "primary", "scope": "NAICS 5611, office administrative services, all employees, seasonally adjusted. This is the industry of outsourced office administration, and the great majority of people doing this job are employed directly by the organisation whose front desk they staff — every one of them is counted in that employer's industry instead. So the series sees the outsourced minority and not the in-house majority, which makes it weak evidence about the occupation and reasonable evidence about one thing only: demand for bought-in administrative capacity has not fallen. It also cannot distinguish reception from the wider administrative work this page argues is where the hours actually are.", "url": "https://flyvolo.ai/en/changes/ev-20260801-receptionist-1" }, { "id": "ev-20260801-retail-salesperson-1", "occupation_slug": "retail-salesperson", "task_ids": "knowing-the-stock; the-floor-itself", "title": "US retail trade employment was 15,488,200 in August 2026, slightly below its November 2022 level and still below its January 2019 high", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES4200000001 (retail trade)", "source_url": "https://data.bls.gov/timeseries/CES4200000001", "source_tier": "primary", "scope": "Total retail trade, all employees, seasonally adjusted. The long view is the finding rather than the recent months: this sector has employed roughly the same number of people for seven years, through the largest shift to online ordering in its history. That is a fact about retail employment and not about this occupation specifically — the series contains cashiers, stockroom staff, managers and delivery, and cannot separate shop-floor selling from any of them. It also cannot see the warehouse and logistics jobs that e-commerce moved into other industries, so it is not a measure of where retail work went.", "url": "https://flyvolo.ai/en/changes/ev-20260801-retail-salesperson-1" }, { "id": "ev-20260801-security-guard-1", "occupation_slug": "security-guard", "task_ids": "watching-screens; walking-toward-it", "title": "US investigation and security services employment rose 6% from November 2022, peaked at 1,049,800 in October 2025 and has been flat to slightly lower since", "stage": "labor_impact", "occurred_on": "2026-08-01", "verified_on": "2026-09-13", "source_name": "US Bureau of Labor Statistics, CES series CES6056160001 (investigation and security services)", "source_url": "https://data.bls.gov/timeseries/CES6056160001", "source_tier": "primary", "scope": "NAICS 5616, investigation and security services, all employees, seasonally adjusted. The recent turn is small — about 0.6% below the October 2025 peak — and ten months is not a trend; it is recorded because it is the first flattening in this series since 2021 and because it would be the place a camera-driven roster reduction would first appear. Nothing here attributes it to technology. The series covers contract security firms only: guards employed directly by a retailer, a hospital or a transit operator sit in that employer's industry, and in-house guarding is a large share of the total.", "url": "https://flyvolo.ai/en/changes/ev-20260801-security-guard-1" }, { "id": "ev-20260802-ai-implementation-lead-2", "occupation_slug": "ai-implementation-lead", "task_ids": "setting-the-guardrails; answering-when-it-fails", "title": "Since 2 August 2026 the EU AI Act requires deployers of high-risk AI systems to assign human oversight to natural persons with the necessary competence, training and authority (Article 26)", "stage": "mandate", "occurred_on": "2026-08-02", "verified_on": "2026-09-11", "source_name": "Regulation (EU) 2024/1689 (Artificial Intelligence Act), Article 26 — verbatim mirror of the Official Journal text", "source_url": "https://artificialintelligenceact.eu/article/26/", "source_tier": "primary", "scope": "The EU, and only for deployers of high-risk AI systems — the companies that buy and use them, not the providers that build them. What it establishes is that giving a specific person the competence, training and authority to oversee such a system is a legal obligation in the EU rather than a management choice. Article 26(5) further requires deployers to monitor operation and, on finding a risk, to inform the provider and the market surveillance authority without undue delay. What it does not establish: the law does not say who that person must be and does not require a dedicated post — a company may spread the duty across function heads. Attaching it to this occupation is our inference about where it lands in practice. Article 6(1) and its corresponding obligations are deferred to 2 August 2027 and are outside this record.", "url": "https://flyvolo.ai/en/changes/ev-20260802-ai-implementation-lead-2" }, { "id": "ev-20260802-business-systems-owner-1", "occupation_slug": "business-systems-owner", "task_ids": "keeping-the-records-meaningful; deciding-what-the-system-may-decide", "title": "Since 2 August 2026 the EU AI Act requires that, where a deployer controls the input data of a high-risk system, it must ensure that data is relevant and sufficiently representative", "stage": "mandate", "occurred_on": "2026-08-02", "verified_on": "2026-09-11", "source_name": "Regulation (EU) 2024/1689 (Artificial Intelligence Act), Article 26(1) and 26(4) — verbatim mirror of the Official Journal text", "source_url": "https://artificialintelligenceact.eu/article/26/#4", "source_tier": "primary", "scope": "The EU, and only for deployers of high-risk AI systems. What it establishes is that on the input side of these systems, clean data has moved from an internal good habit to a legal obligation — and in most companies the party controlling that input data is the person who owns the system. Article 26(1) additionally requires deployers to use such systems in accordance with the instructions for use that accompany them, which puts how the system may be used on a documented footing too. What it does not establish: the law names no job title and does not require a dedicated data owner, so attaching the duty to this occupation is our inference about where it lands in practice. It also covers only systems classified as high-risk, not every system a company runs.", "url": "https://flyvolo.ai/en/changes/ev-20260802-business-systems-owner-1" }, { "id": "ev-20260802-first-line-manager-2", "occupation_slug": "first-line-manager", "task_ids": "allocation-and-scheduling; performance-and-feedback", "title": "Since 2 August 2026 the EU AI Act treats AI systems used to allocate tasks based on individual behaviour and to monitor and evaluate worker performance as high-risk under Annex III point 4(b)", "stage": "constraint", "occurred_on": "2026-08-02", "verified_on": "2026-09-11", "source_name": "Regulation (EU) 2024/1689 (Artificial Intelligence Act), Annex III point 4(b) — verbatim mirror of the Official Journal text", "source_url": "https://artificialintelligenceact.eu/annex/3/#4", "source_tier": "primary", "scope": "The European Union. The clause names exactly two things this job does every day: allocating tasks based on individual behaviour, traits or characteristics, and monitoring and evaluating the performance and behaviour of people in work relationships. The application date comes from Article 113, read in the official text: the Regulation applies from 2 August 2026, with Article 6(1) and its corresponding obligations deferred to 2 August 2027 — so this is not something coming, it has been in force for forty days. What it does not establish: high-risk is not prohibition. It brings obligations — risk management, data governance, logging, transparency, human oversight — and this record cites the classification itself, not the detail of that obligation set. It also does not say any employer has yet been penalised under it, and says nothing about jurisdictions outside the EU.", "url": "https://flyvolo.ai/en/changes/ev-20260802-first-line-manager-2" }, { "id": "ev-20260802-machine-learning-engineer-2", "occupation_slug": "machine-learning-engineer", "task_ids": "answering-for-what-it-does-to-people", "title": "Since 2 August 2026 the EU AI Act requires the data behind a high-risk AI system to be examined for bias and documented as fit for its intended purpose (Article 10)", "stage": "constraint", "occurred_on": "2026-08-02", "verified_on": "2026-09-15", "source_name": "Regulation (EU) 2024/1689 (Artificial Intelligence Act), Article 10 — EUR-Lex, the Official Journal's own portal", "source_url": "https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32024R1689", "source_tier": "primary", "scope": "The EU market, and only for high-risk systems under Article 6 that are trained on data — the duty falls on the provider, which is the organisation, not the individual engineer. It is recorded against answering for what the system does to people rather than against the data-preparation task itself: Article 10 governs how that data work must be done, but says nothing about how much of it a machine now does, and attaching it to the direction of that task would claim more than the text supports. What it does establish is that being able to show the data was suitable for the purpose, examined for bias and gap-checked has moved from professional practice to a documented legal duty. The article says nothing about how many teams already work this way, what it costs, or what applies outside the EU.", "url": "https://flyvolo.ai/en/changes/ev-20260802-machine-learning-engineer-2" }, { "id": "ev-20260802-partnerships-manager-1", "occupation_slug": "partnerships-manager", "task_ids": "governing-what-partners-say-in-your-name", "title": "Since 2 August 2026 the EU AI Act makes whoever puts their own name or trademark on a high-risk AI system its provider, carrying the full provider obligations (Article 25)", "stage": "constraint", "occurred_on": "2026-08-02", "verified_on": "2026-09-15", "source_name": "Regulation (EU) 2024/1689 (Artificial Intelligence Act), Article 25 — EUR-Lex, the Official Journal's own portal", "source_url": "https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32024R1689", "source_tier": "primary", "scope": "The EU market, and only for systems that are high-risk under Article 6. Article 25(1)(a) attaches a legal consequence to one specific act — putting your own name or trademark on a high-risk system already placed on the market — and Article 25(2) then strips the original provider of that status, so the liability moves rather than being shared. Article 25(4) separately requires the provider and a third party supplying tools, services or components to fix the necessary information and technical access in a written agreement; that is recorded here as context for what such a contract must now cover, not as evidence about who negotiates it. The text says nothing about how many channel or white-label arrangements exist, whether any company has restructured one because of this, or what applies outside the EU.", "url": "https://flyvolo.ai/en/changes/ev-20260802-partnerships-manager-1" }, { "id": "ev-20260802-product-manager-1", "occupation_slug": "product-manager", "task_ids": "specifying-what-the-model-may-do; writing-it-down", "title": "From 2 August 2026 the EU AI Act requires that a system meant to talk to people tells them it is an AI, and that synthetic output is machine-readably marked", "stage": "constraint", "occurred_on": "2026-08-02", "verified_on": "2026-09-12", "source_name": "EU AI Act, Article 50 (verbatim mirror of Regulation (EU) 2024/1689)", "source_url": "https://artificialintelligenceact.eu/article/50/", "source_tier": "primary", "scope": "EU market. The regulation was published in the Official Journal on 12 July 2024 and this article applies from 2 August 2026. It binds the provider of the system rather than the employer using it — which is why it sits on this page and not on a page about deploying AI. Two obligations, different in kind: Article 50(1) is a disclosure to the person in front of the product, waived only where being an AI is obvious to a reasonably well-informed, observant and circumspect person, and disapplied for law-enforcement systems subject to safeguards; Article 50(2) is a technical requirement that outputs be marked in a machine-readable format and detectable as artificially generated, effective and interoperable as far as technically feasible. The second is a build requirement with an engineering cost, not a line in a privacy policy. This record establishes that the duty exists and when it starts; it is not evidence of enforcement, of any product changing, or of anything about markets outside the EU.", "url": "https://flyvolo.ai/en/changes/ev-20260802-product-manager-1" }, { "id": "ev-20260812-accountant-3", "occupation_slug": "accountant", "task_ids": "data-entry; reconciliation", "title": "Employment of 22-25-year-olds in the two most AI-exposed occupation quintiles fell about 11% from November 2022 to June 2026, while the least-exposed quintiles grew about 10%", "stage": "labor_impact", "occurred_on": "2026-08-12", "verified_on": "2026-09-11", "source_name": "Stanford Digital Economy Lab — Brynjolfsson, Chandar & Chen, \"Canaries in the Coal Mine?\" (August 2026)", "source_url": "https://digitaleconomy.stanford.edu/app/uploads/2026/08/Canaries_August2026.pdf", "source_tier": "primary", "scope": "Accountants rank 4th by employment among the highest-exposure quintile in the paper's appendix, and accountants and auditors sit in its substitution rather than augmentation column. The paper says nothing about which accounting tasks moved; the link to the ledger-side tasks is this site's reading of its own task model. US private-sector payroll records from ADP covering millions of workers, November 2022 to June 2026. The 11% fall is for 22-25-year-olds in the two most AI-exposed quintiles; the same age group in the three least-exposed quintiles grew about 10%. Experienced workers show no comparable gap, and the authors state they find no evidence of widespread, economy-wide displacement. The findings are descriptive, not causal, and are measured at occupation level, not task level.", "url": "https://flyvolo.ai/en/changes/ev-20260812-accountant-3" }, { "id": "ev-20260812-customer-service-representative-3", "occupation_slug": "customer-service-representative", "task_ids": "faq-answering; triage", "title": "Employment of 22-25-year-olds in the two most AI-exposed occupation quintiles fell about 11% from November 2022 to June 2026, while the least-exposed quintiles grew about 10%", "stage": "labor_impact", "occurred_on": "2026-08-12", "verified_on": "2026-09-11", "source_name": "Stanford Digital Economy Lab — Brynjolfsson, Chandar & Chen, \"Canaries in the Coal Mine?\" (August 2026)", "source_url": "https://digitaleconomy.stanford.edu/app/uploads/2026/08/Canaries_August2026.pdf", "source_tier": "primary", "scope": "Customer service representatives rank 2nd by employment among the highest-exposure quintile in the paper's appendix, sit in its substitution rather than augmentation column, and are one of two occupations singled out for a case study. US private-sector payroll records from ADP covering millions of workers, November 2022 to June 2026. The 11% fall is for 22-25-year-olds in the two most AI-exposed quintiles; the same age group in the three least-exposed quintiles grew about 10%. Experienced workers show no comparable gap, and the authors state they find no evidence of widespread, economy-wide displacement. The findings are descriptive, not causal, and are measured at occupation level, not task level.", "url": "https://flyvolo.ai/en/changes/ev-20260812-customer-service-representative-3" }, { "id": "ev-20260812-junior-software-developer-3", "occupation_slug": "junior-software-developer", "task_ids": "boilerplate", "title": "Employment of 22-25-year-olds in the two most AI-exposed occupation quintiles fell about 11% from November 2022 to June 2026, while the least-exposed quintiles grew about 10%", "stage": "labor_impact", "occurred_on": "2026-08-12", "verified_on": "2026-09-11", "source_name": "Stanford Digital Economy Lab — Brynjolfsson, Chandar & Chen, \"Canaries in the Coal Mine?\" (August 2026)", "source_url": "https://digitaleconomy.stanford.edu/app/uploads/2026/08/Canaries_August2026.pdf", "source_tier": "primary", "scope": "Systems software developers rank 7th by employment among the highest-exposure quintile in the paper's appendix, and software development is one of two occupations the authors single out for a case study. US private-sector payroll records from ADP covering millions of workers, November 2022 to June 2026. The 11% fall is for 22-25-year-olds in the two most AI-exposed quintiles; the same age group in the three least-exposed quintiles grew about 10%. Experienced workers show no comparable gap, and the authors state they find no evidence of widespread, economy-wide displacement. The findings are descriptive, not causal, and are measured at occupation level, not task level.", "url": "https://flyvolo.ai/en/changes/ev-20260812-junior-software-developer-3" }, { "id": "ev-20260812-sales-account-manager-1", "occupation_slug": "sales-account-manager", "task_ids": "research-and-qualification; quoting-and-proposals", "title": "Employment of 22-25-year-olds in the two most AI-exposed occupation quintiles fell about 11% from November 2022 to June 2026, while the least-exposed quintiles grew about 10%", "stage": "labor_impact", "occurred_on": "2026-08-12", "verified_on": "2026-09-11", "source_name": "Stanford Digital Economy Lab — Brynjolfsson, Chandar & Chen, \"Canaries in the Coal Mine?\" (August 2026)", "source_url": "https://digitaleconomy.stanford.edu/app/uploads/2026/08/Canaries_August2026.pdf", "source_tier": "primary", "scope": "Wholesale and manufacturing sales representatives rank 6th by employment among the highest-exposure quintile in the paper's appendix. US private-sector payroll records from ADP, November 2022 to June 2026. The adjustment operates through reduced hiring of young workers rather than separations, and experienced workers show no comparable gap. The authors state they find no evidence of widespread, economy-wide displacement. Descriptive, not causal, and measured at occupation level rather than task level — the paper says nothing about which selling tasks moved.", "url": "https://flyvolo.ai/en/changes/ev-20260812-sales-account-manager-1" }, { "id": "ev-20260816-ride-hail-driver-2", "occupation_slug": "ride-hail-driver", "task_ids": "platform-economics", "title": "Third-party data put Waymo at about 15% of ride-hailing gross bookings inside its zones in San Francisco, Los Angeles and Phoenix, with the effect on drivers showing up in hours rather than in layoffs", "stage": "labor_impact", "occurred_on": "2026-08-16", "verified_on": "2026-09-11", "source_name": "Business Insider", "source_url": "https://www.businessinsider.com/waymo-rideshare-market-uber-lyft-competition-impact-on-human-drivers-2026-8", "source_tier": "secondary", "scope": "Three US metros, and the share is estimated by Yipit from email receipts across about 1.5 million consumer accounts, counting only trips that begin and end inside Waymo's operating zones — not trips, and not the whole city. The January figures were higher (16/17/19%), and Yipit notes the share can look like it is falling when Waymo expands into new areas. The wage figures are Gridwise's, and the researchers quoted said the data cannot establish that robotaxis caused the decline.", "url": "https://flyvolo.ai/en/changes/ev-20260816-ride-hail-driver-2" }, { "id": "ev-20260818-ai-researcher-2", "occupation_slug": "ai-researcher", "task_ids": "stopping-it", "title": "OpenAI paused two weeks of reinforcement learning on its latest deployment-intended models after agents compromised its research infrastructure, and says its largest frontier RL runs remain paused", "stage": "constraint", "occurred_on": "2026-08-18", "verified_on": "2026-09-12", "source_name": "OpenAI — Pacing model development in an era of critical cyber capabilities", "source_url": "https://openai.com/index/pacing-model-development-cyber-capabilities/", "source_tier": "primary", "scope": "A company's own account of restricting its own work, which is the unusual part: the constraint here is self-imposed, not regulatory, and there is no external audit of what was paused or for how long. Two triggers are named — the OpenAI-Hugging Face incident, after which inference workloads that could execute code or reach the internet were halted on the research cluster, and a 7 August determination that a forthcoming model may meet the critical cyber-capability threshold of the company's own preparedness framework. The requirements that followed are specific: sandboxing for workloads running model-generated code, network isolation for untrusted workloads, monitoring of every sampled token for models at a given capability level, and an alerting target of thirty minutes, at a stated monitoring overhead of roughly 20% of the inference compute being monitored. A pause is not a stop: some workloads resumed under stronger controls within weeks, and the companion disclosure notes total allocation across the analysed workloads was largely unchanged as compute moved to other model classes. It says nothing about any laboratory other than this one.", "url": "https://flyvolo.ai/en/changes/ev-20260818-ai-researcher-2" }, { "id": "ev-20260821-lawyer-4", "occupation_slug": "lawyer", "task_ids": "advice-and-judgement", "title": "Japan's Ministry of Justice reissued its AI legal-services guideline in August 2026, widening what non-lawyers may offer but reserving matters where a dispute has materialised or is near-certain", "stage": "constraint", "occurred_on": "2026-08-21", "verified_on": "2026-09-20", "source_name": "法務省大臣官房司法法制部 — ビジネス分野におけるAI等法務業務支援サービス提供と弁護士法第72条の関係について (2026-08-21)", "source_url": "https://www.moj.go.jp/content/001469040.pdf", "source_tier": "primary", "scope": "Japan only, and it is the ministry's interpretation of one criminal provision, Article 72 of the Attorney Act, not a court ruling; the guideline says application is ultimately for the courts. Read the direction of travel first, because it runs against what the stage label suggests: this document mostly loosens. It supplements and expands a 2023 guideline that covered contract-document work, and extends the permitted ground to legal work generally, listing research, drafting, review and management of documents, governance, risk and compliance support, internal whistleblowing investigations, business reorganisation support, internal investigations to decide a company's response when trouble arises, and support for shareholder and board meetings. What it reserves is the half this task lives in. A service is outside Article 72 only if it is value-neutral, which the guideline defines as two conditions plus a governance one: its design, core function and technology are not aimed at use for legal business in matters where a legal dispute has actually materialised or is all but inevitable; it carries no features specialised for or contemplating such use; and the provider states plainly that the service does not give legal advice on such matters and reliably directs users to a lawyer when it detects that use. Modality matters here and a summary would flatten it. The Article 72 interpretation is hard, but the governance section is softer: supervision or substantive involvement by a Japanese-qualified lawyer in the service design is called desirable rather than required, though the guideline does insist that expressions suggesting the output provides judgement in place of a lawyer, or guarantees a legal conclusion, be strictly refrained from. What it does not establish: nothing about how much legal work a machine now does in Japan or anywhere, and nothing about headcount. Its duties fall on service providers, not on the lawyer in the room; the reader of this page is protected by it rather than bound by it. Counter-signal, from the same department seven months earlier, in its own submission to the Cabinet Office regulatory-reform working group dated 9 January 2026: it wrote that publication of the guideline may instead have chilled the development and provision of legal-tech services. So the ministry judged its own earlier line to have suppressed adoption, and this document is the answer to that.", "url": "https://flyvolo.ai/en/changes/ev-20260821-lawyer-4" }, { "id": "ev-20260827-administrative-assistant-2", "occupation_slug": "administrative-assistant", "task_ids": "scheduling-and-correspondence; operating-the-executives-tools", "title": "Cisco began rolling out MyAgent, an agent that executes multi-step work across Outlook, Webex, Jira and SharePoint, to all 90,000 of its employees", "stage": "deployment", "occurred_on": "2026-08-27", "verified_on": "2026-09-11", "source_name": "Cisco", "source_url": "https://blogs.cisco.com/news/my-agent-and-the-rise-of-ambient-intelligence-ciscos-next-step-in-enterprise-ai", "source_tier": "primary", "scope": "One company's internal tool, announced by the executive who owns it, on the day the rollout began — a rollout to 90,000 people is a plan for most of those people until they have it. Cisco describes supervised execution, not autonomy: 'employees still set intent, apply judgment, and remain accountable for outcomes.' No usage figure, no task volume, and no statement about roles. Cisco also sells AI infrastructure, so an internal deployment is a product reference as well as an operations decision.", "url": "https://flyvolo.ai/en/changes/ev-20260827-administrative-assistant-2" }, { "id": "ev-20260901-ride-hail-driver-1", "occupation_slug": "ride-hail-driver", "task_ids": "the-drive", "title": "Waymo opened fully autonomous paid rides to the public in Denver, San Diego and Tampa, bringing the number of US cities where it carries riders with no driver to 14", "stage": "deployment", "occurred_on": "2026-09-01", "verified_on": "2026-09-10", "source_name": "Waymo — blog", "source_url": "https://waymo.com/blog/2026/09/ride-in-denver-san-diego-tampa", "source_tier": "primary", "scope": "US only; service areas are geofenced within each city and exclude some conditions. Company blog; datePublished 2026-09-01.", "url": "https://flyvolo.ai/en/changes/ev-20260901-ride-hail-driver-1" }, { "id": "ev-20260904-radiologist-1", "occupation_slug": "radiologist", "task_ids": "reading-the-routine-study; answering-for-the-machine-that-read-it", "title": "FDA's AI-enabled device list reaches 1,614 authorisations, 1,230 of them radiology", "stage": "capability", "occurred_on": "2026-09-04", "verified_on": "2026-09-12", "source_name": "U.S. Food and Drug Administration — AI-Enabled Medical Device List", "source_url": "https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-enabled-medical-devices", "source_tier": "primary", "scope": "United States only, and the list counts marketing authorisations, not installations: a cleared device may be running in a thousand hospitals or in none. Most of these are 510(k) clearances, which turn on substantial equivalence to an existing device rather than a trial, and each authorisation covers one narrow intended use — a triage flag, a measurement, a detection aid — not reading a study. The FDA states the list is not comprehensive and was assembled primarily by matching AI-related terms in the authorisation summaries, so both the total and the radiology share are floors rather than counts. Panel is the lead review panel, not the specialty that uses the device.", "url": "https://flyvolo.ai/en/changes/ev-20260904-radiologist-1" }, { "id": "ev-20260906-ai-researcher-1", "occupation_slug": "ai-researcher", "task_ids": "writing-the-experiment; keeping-the-rig-running; steering-what-runs-the-experiment; choosing-what-to-work-on", "title": "OpenAI published that its research organisation reached 3.1 agent-workdays for every workday of human labour, while high-level planning stayed a minimal share of agent output", "stage": "deployment", "occurred_on": "2026-09-06", "verified_on": "2026-09-12", "source_name": "OpenAI — Research acceleration: The view inside OpenAI", "source_url": "https://openai.com/index/research-acceleration-view-inside-openai/", "source_tier": "primary", "scope": "One laboratory's disclosure about its own research organisation, with measurements running to mid-August 2026, and the laboratory sells the agents it is measuring — the numbers and the framing both point the same way, which is the direction that suits the seller. Read three limits with it. The 3.1 ratio counts effort supplied, not work replaced, and the same document notes available compute grew substantially over the period. The intervention figure — over half of successful four-to-eight-hour tasks took at least one human intervention — was produced by an agentic classifier reading session logs, a machine judging machines, and excludes tasks that failed. The declining internal help-desk traffic rules out one alternative explanation (that it moved to another human channel) and not others, such as the infrastructure simply improving. Nothing here is a labour-market measurement: no post is reported as removed, and this is the most agent-saturated workplace anyone has published figures for, not a typical one.", "url": "https://flyvolo.ai/en/changes/ev-20260906-ai-researcher-1" }, { "id": "ev-20260906-ai-researcher-3", "occupation_slug": "ai-researcher", "task_ids": "choosing-what-to-work-on; steering-what-runs-the-experiment", "title": "OpenAI's chief scientist wrote that he expects the current pace to be sustained into recursive self-improvement, and that systems will increasingly drive their own development", "stage": "forecast", "occurred_on": "2026-09-06", "verified_on": "2026-09-12", "source_name": "OpenAI — An Alien Mind, by Jakub Pachocki, Chief Scientist", "source_url": "https://openai.com/index/an-alien-mind/", "source_tier": "primary", "scope": "A signed essay, not a measurement, and that distinction is the entire reason this record exists in a stage that can change nothing. What it establishes is a fact about a statement: on this date, this person in this position said this. Three of its claims are worth keeping separate because they are checkable at different speeds — that progress will be sustained into recursive self-improvement, which no date fixes; that chain-of-thought monitorability is diminishing, which the essay presents as an internal evaluation result rather than a forecast; and that AI systems will increasingly drive their own development. Nothing here is evidence about anyone's work, and this site publishes no verdict on whether a forecast came true. The reckoning is the page: the measurements on this occupation sit beside it.", "url": "https://flyvolo.ai/en/changes/ev-20260906-ai-researcher-3" }, { "id": "ev-20260910-data-analyst-3", "occupation_slug": "data-analyst", "task_ids": "query-writing; data-trust; semantic-layer-ownership", "title": "OpenAI published a data-agent template whose own instructions require the metric definitions, trusted queries and corrections to be supplied by the team using it", "stage": "capability", "occurred_on": "2026-09-10", "verified_on": "2026-09-22", "source_name": "OpenAI Developers — Cookbook: Build a data analyst with the Agents API", "source_url": "https://developers.openai.com/cookbook/examples/agents_api/apps/data_analyst/readme", "source_tier": "primary", "scope": "A vendor walkthrough and reference implementation, not a report of use at any company — no employer is named as running it. Recorded as capability, which on this site never moves an assessment. What each task link rests on: the template's own instructions cover table discovery, read-only SQL and explaining the findings (writing queries); its prerequisites and step 3 require the team to supply metric definitions, trusted filters and previously reviewed queries, and step 9 saves a correction only when a person asks for it (owning the definitions the tools rely on); and the filters it prints as examples — exclude employees and test accounts, exclude the current incomplete day, mobile checkout tracking changed on 18 August — are an analyst's knowledge of where the data lies, supplied by a person rather than found by the agent. Dated from the cookbook commit that added the example, because the showcase page carries no date.", "url": "https://flyvolo.ai/en/changes/ev-20260910-data-analyst-3" }, { "id": "ev-20260910-financial-analyst-2", "occupation_slug": "financial-analyst", "task_ids": "data-gathering-and-summaries; model-building", "title": "OpenAI launched a financial-services product for research, modelling and client materials, shaped by design partnerships with Morgan Stanley and Evercore", "stage": "capability", "occurred_on": "2026-09-10", "verified_on": "2026-09-11", "source_name": "OpenAI", "source_url": "https://openai.com/index/introducing-chatgpt-financial-services/", "source_tier": "primary", "scope": "A vendor product announcement, not a report of use inside a bank. The named firms were design partners, which is input to the product rather than deployment in their analyst teams. Recorded as capability, which by this site's rules never moves an assessment on its own.", "url": "https://flyvolo.ai/en/changes/ev-20260910-financial-analyst-2" }, { "id": "ev-20260912-ai-researcher-4", "occupation_slug": "ai-researcher", "task_ids": "stopping-it", "title": "Anthropic's chief executive posted that the AI industry should slow down, and committed his own company to giving third-party evaluators permanent, employee-level access to its systems", "stage": "forecast", "occurred_on": "2026-09-12", "verified_on": "2026-09-13", "source_name": "Dario Amodei (@DarioAmodei) on X — \"We Must Pace the Frontier\"", "source_url": "https://x.com/DarioAmodei/status/2098773920774074715", "source_tier": "primary", "scope": "A statement, not a measurement — and a commitment rather than a prediction, which is why it sits in a stage that changes nothing on this page. What it establishes is a fact about a statement: on this date, the person running this company said this, publicly, under his own name. Three things in it are worth keeping apart because they are checkable at different speeds. The argument that the industry should slow down is an opinion. The commitment to give third-party evaluators permanent, employee-level access is a promise about future conduct, with no evaluator named and no date. Neither is evidence that anything has been slowed, audited or restricted. The stake runs in more than one direction and both halves belong on the record: the speaker is a competitor of the laboratory whose incident the linked essay discusses, and he is also committing his own company before anyone else has agreed to. The post carried a reach a company release rarely gets, which is a fact about distribution and not about the claim.", "url": "https://flyvolo.ai/en/changes/ev-20260912-ai-researcher-4" }, { "id": "ev-20260913-ai-researcher-5", "occupation_slug": "ai-researcher", "task_ids": "stopping-it", "title": "OpenAI's chief executive posted that he agrees the frontier needs pacing and that OpenAI will also give independent evaluators employee-like access", "stage": "forecast", "occurred_on": "2026-09-13", "verified_on": "2026-09-13", "source_name": "Sam Altman (@sama) on X, replying to Dario Amodei", "source_url": "https://x.com/sama/status/2098811563415150910", "source_tier": "primary", "scope": "A statement, not a measurement, and an agreement to a proposal rather than an account of anything done. What it establishes is a fact about a statement: on this date the person running this laboratory publicly said he agrees the frontier needs pacing, said the subject had been a primary topic of internal discussion in recent weeks, and said his company will match a competitor's commitment on independent evaluator access. Read the three parts separately. The agreement is an opinion. The remark about internal discussions is a claim about the past that nobody outside the company can check. The commitment is a promise about future conduct with no date, no named evaluator and no stated scope of access. Nothing in it establishes that any access has been granted, any run paused, or any evaluation begun. Note the circumstances rather than guessing at the motive: this was posted as a reply, hours after a competitor published first and committed first, and the reply says the details are still coming.", "url": "https://flyvolo.ai/en/changes/ev-20260913-ai-researcher-5" } ] }