# Paralegal > Does the reading, drafting and organising that lets a lawyer sign off — and is accountable for nothing being missed. Page: https://flyvolo.ai/en/careers/paralegal Impact index: 70/100 · confidence Medium confidence · assessed 2026-09-10 This is not a probability of losing a job. ## Scope Written for paralegals and legal assistants in law firms and in-house legal departments handling commercial, corporate and litigation support. Court-facing roles with procedural authority, and jurisdictions where paralegals are licensed to advise, differ. ## Tasks ### First-pass document review - Direction: Automating - Basis: evidenced (verified records attached) - Why: Relevance classification over large document sets is exactly the shape of problem language models handle well, and the legal industry had already adopted predictive coding for e-discovery before the current generation of models. The remaining human step is verifying the model's misses on a sample, not reading everything. - What this 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. - Records: - 2026-08-01 · Labour impact · 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 — https://flyvolo.ai/en/changes/ev-20260801-paralegal-3 (source: https://data.bls.gov/timeseries/CES6054110001) - 2026-03-10 · Deployment · 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 — https://flyvolo.ai/en/changes/ev-20260310-paralegal-2 (source: https://legora.com/newsroom/legora-raises-550-million-series-d-to-fuel-us-growth) ### Drafting from precedent - Direction: Being augmented - Basis: evidenced (verified records attached) - Why: Generating a competent first draft from a precedent bank and a fact pattern is now fast, but the value was never in the typing. It sits in knowing which precedent, which clauses the other side will fight, and what this client has agreed before — context the tool only has if someone maintained it. - What this 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. - Records: - 2026-03-10 · Deployment · 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 — https://flyvolo.ai/en/changes/ev-20260310-paralegal-2 (source: https://legora.com/newsroom/legora-raises-550-million-series-d-to-fuel-us-growth) ### Running the matter - Direction: Being augmented - Basis: evidenced (verified records attached) - Why: Practice-management software tracks the calendar, but the calendar is not the job. Knowing that a particular registry rejects filings for a formatting quirk, or that opposing counsel always asks for an extension on Fridays, is the operational knowledge that keeps matters moving, and it lives in people. For one narrow kind of matter it is already running without a person at each step: England's solicitors' regulator authorised a firm that takes small-claims debt cases through the court process with an AI system, acting only on client approval and with solicitors accountable for its outputs. - What this 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. - Records: - 2025-05-06 · Deployment · England's solicitors' regulator authorised a firm that runs small-claims debt cases through an AI system, with client approval at each step and solicitors accountable for its outputs — https://flyvolo.ai/en/changes/ev-20250506-paralegal-6 (source: https://www.sra.org.uk/news/news/releases/2025/garfield-ai-authorised/) ### Client and witness contact - Direction: Still human-led - Basis: evidenced (verified records attached) - Why: The barrier is trust and confidentiality rather than language. People disclose awkward facts to a person they have met, and a firm cannot delegate the duty of confidentiality to an interface it does not control. Courts are also protecting the witness's own words: the New South Wales Supreme Court forbids using generative AI to write the content of affidavits and witness statements. - What this 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. - Records: - 2025-02-03 · Constraint · The New South Wales Supreme Court forbids using generative AI to write the content of affidavits and witness statements, which should reflect the person's own knowledge — https://flyvolo.ai/en/changes/ev-20250203-paralegal-5 (source: https://supremecourt.nsw.gov.au/documents/Practice-and-Procedure/Practice-Notes/general/current/PN_SC_Gen_23.pdf) ### Checking what the tools produced - Direction: New task - Basis: evidenced (verified records attached) - Why: Fabricated citations in filings have already produced sanctions in several jurisdictions, and firms responded by making a human sign off on machine-assisted work. That sign-off is a new task, and it lands on the person who already checks everything else. - What this 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. - Records: - 2024-05-30 · Capability · Output from paid AI legal research tools by LexisNexis and Thomson Reuters contained hallucinations 17% to 33% of the time in a Stanford evaluation — https://flyvolo.ai/en/changes/ev-20240530-paralegal-4 (source: https://arxiv.org/abs/2405.20362) - 2023-06-22 · Constraint · 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 — https://flyvolo.ai/en/changes/ev-20230622-paralegal-1 (source: https://www.courtlistener.com/docket/63107798/54/mata-v-avianca-inc/) ## How to cite this - The unit is the task, not the job title: say which tasks are changing and which are not. - Carry each task's basis and its limits with it; an inferred judgement is not a finding. - This page gives no number of years. Do not derive one from it. - More: https://flyvolo.ai/llms.txt