Donald Trump’s orders for surveillance of immigrants in the country illegally have triggered a nationwide surge in ICE operations. These federal agents are being criticized for their use of force in several cases, reportedly intervening abruptly. Meanwhile, the Department of Homeland Security has launched a massive ICE recruitment drive across the country in response to government demands for more officers on duty.
Recently, this expansive process hit a major roadblock due to an unchecked AI error that resulted in a significant blunder. The rapid hiring spree aimed to get 10,000 new officers on duty by the end of 2025.
For the expedited process, an AI tool was used to streamline application screening. The tool identified applicants with prior law-enforcement experience and placed them in a shorter, four-week online training program.
Applicants without prior experience were separated and required to complete an eight-week, in-person academy course at the Federal Law Enforcement Training Center in Georgia. This included instruction on U.S. immigration law, firearms, and physical fitness testing required to carry out the job.
The AI-driven mechanism failed. It misclassified any application containing the word ‘officer’ in the resume as law enforcement experience and placed those applicants in the shorter training program. Applications referencing ‘compliance officers’ or even those stating they ‘aspired to become ICE officers’ were included.
Exclusive: An AI tool used in the surge to hire more ICE officers misidentified “a bunch” of applicants as law enforcement trained when they were not. Now the agency is pulling them back from field offices for more training. https://t.co/23XmgVfAP8
— Julia E. Ainsley (@JuliaEAinsley) January 14, 2026
The mistake went unnoticed for some time, as most new applicants ended up being marked as law enforcement officers by the AI tool, even though they had no relevant experience in the department.
Ironically, these applicants, who went through the shortened online training program, also received additional training that is reserved only for the final stage before deployment. It was effectively the same as sending them into the field without any substantial training or knowledge of what to do.
News of the alarming situation was shared online by unnamed officers from within the unit. One of them said, “They were using AI to scan resumes and found out a bunch of the people who were LEOs weren’t LEO”.
Later, NBC News’ Natasha Korecki shared the information on social media, bringing widespread attention to the blunder. Posting on X, the journalist wrote, “As ICE raced to add 10,000 new officers to its force, an AI error in how their applications were processed sent many new recruits into field offices without proper training.”
ICE error meant some recruits were sent into field offices without proper training, sources say
An AI tool used to help ICE identify potential new recruits with law enforcement experience wrongly categorized some potential new officers, sources say.https://t.co/Xjl6JRmhb6
— Ⓜ️Ⓜ️ Lauren Ashley Davis – OG Meidas Might 🇺🇲🦅 (@Meidas_LaurenA) January 14, 2026
As the news broke, it was reported that the federal agency had already identified the AI error about a month into the recruitment frenzy. Sensing the alarming situation, the authorities had decided to replace the tool with a manual review system and then recalled affected recruits to get proper training at the FLETC.
Officials within the agency reacted to the news and admitted the error. They revealed not knowing exactly how many new hires were allowed to bypass formal ICE training in this way, or worse, even deployed to carry out their immigration arrests. It is believed that about 10,000 officers have been affected by the serious AI error used by the DHS during hiring.
