LO
DeploymentCognitive automation2025-04-15
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
Experienced software engineeroccupation page →
Event date / reported
2025-04-15 · reported 2026-01-28
Evidence stage
DeploymentAn employer has put it into production. Can move the baseline — weighted by scale and how similar the setting is.
Tasks this bears on
Reviewing what the machine wrote
Reading code you did not write and were not there for, and deciding whether it is right — at a volume that rises as generation gets cheaper.
New task✓ Evidence-backed
Owning the agents that write and change code
Setting what an automated coding agent may do unattended, what it must ask about, and where its output enters the branch — then being the name attached to that setting.
New task✓ Evidence-backed
Where this applies
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.
What this means
An employer put these tools into production and, in the same filing, wrote down where their output is allowed to go. That boundary is the job: someone had to decide that nothing enters the branch unreviewed and nothing touches production directly, and someone has to keep deciding it as the tools get better at the part that used to justify the rule.
What it does not yet show
It says nothing about volume. The inventory has no figure for how many changes these tools produced, how many were accepted, or how much time anyone saved, and the productivity claim in it is the agency's own sentence with nothing behind it. It also says nothing about whether the review is real: a rule that output must be approved and a reviewer who actually reads it are different things, and no inventory field can tell them apart.
What you can check
Find out whether your own team has written this boundary down anywhere — what an agent may change unattended, and where its output enters the branch. If it exists only as a shared understanding, write one sentence of it and send it to the team; the disagreement you get back is the map of where the boundary actually is.
Does it change the assessment?
No. The impact index is never moved by a single event. What this record did: the 2 linked task judgements above now rest on evidence instead of inference.
Source
OMB — 2025 Federal Agency AI Use Case Inventory (entry DHS-2758) · verified 2026-09-20 · Claude (agent) · interpreted 2026-09-20 · Claude (agent)
Primary source — published by the party that did this, or the authority of record. No co-signature needed.