AIFoPa-2026-0024 — It Looked for the Word Officer

Through the second half of 2025 United States Immigration and Customs Enforcement was under a mandate. It was to add ten thousand officers to its force by the end of the year. Signing bonuses of fifty thousand dollars were drawn on the appropriation made under the One Big Beautiful Bill.

Applications arrived at a rate no established process had been built for. To sort them, the agency put an artificial intelligence tool on the résumés.

The tool’s function was narrow and consequential. It determined which applicants were already law enforcement officers and which were not.

The LEO programme is the short course for applicants who are already law enforcement officers. It runs four weeks, online. Those who were not belonged at the Federal Law Enforcement Training Center in Georgia, for eight weeks in person. The course covers immigration law, the handling of a firearm, and a physical fitness assessment.

The eight-week course is itself the shortened form. It ran to twenty weeks until it was reduced, in the agency’s published account, to cut redundancy and to incorporate advances in technology.

The tool looked for the word officer. It found the word officer in a number of places.

A compliance officer was an officer. So was an applicant who had written that he aspired to become an ICE officer. Two law enforcement officials described this to NBC News. The instrument read a statement of intention as a statement of history.

The majority of new applicants, those officials said, were categorised as serving law enforcement officers. Many had never served in any police force, local or federal.

One of them put the finding in a sentence the Bureau considers exemplary in its economy. “They were using AI to scan résumés and found out a bunch of the people who were LEOs weren’t LEOs.”

The error was identified in mid-autumn, rather more than a month into the recruitment surge. The agency introduced a corrective measure. It began reading the résumés by hand.

Recruits who had been misrouted were recalled to Georgia for the full course. The officials, who were not authorised to speak publicly, told NBC News that they could not establish how many people had been improperly trained. Nor could they establish how many had been sent out before the mistake was found.

The Department of Homeland Security is on record and its account differs. A departmental spokesperson described the matter as a “technological snag” which had been quickly rectified. The spokesperson put the number affected at about two hundred hires. All of them, the spokesperson said, subsequently reported to the Federal Law Enforcement Training Center for full training.

No candidate, the spokesperson said, was placed on enforcement duties without “appropriate training and credentials”. “The vast majority of new officers brought on during the hiring surge are experienced law enforcement officers who have already successfully completed a law enforcement academy.”

The departmental figure and the officials’ account are set down beside one another, in that order. No reconciliation is offered, there being none. One of the two parties has a number, and a number is not the same as no number.

There is a further finding, and it concerns arithmetic rather than training. One official observed that the agency met its target on paper: ten thousand hires, recorded, by the end of 2025.

Recalling the misidentified recruits to be trained properly meant that ten thousand officers were not in fact added to the street in that year.

A count was produced. The count was accurate as to the documents. The documents had been classified by a system. It was measuring something adjacent to the thing the count was understood to be counting. It was not measuring the same thing.

The Bureau declines to record this as a malfunction. On the evidence available the instrument did not malfunction. It was asked to find a word and it found the word, thoroughly and at scale. It did so without hesitating once. A person performing the same task would have hesitated somewhere around the fourth compliance officer.

What was delegated was not the search. What was delegated was the judgement of whether a stranger’s prior life had prepared him for four weeks of instruction or eight. That judgement was rendered on the presence of a substring.

The incident is filed under the classification created for it. The classification will not be lonely for long.