Judge Rips Immigration Agent for Using ChatGPT to Write Protest Report

A federal judge called out federal officers in Chicago for producing unreliable use-of-force reports—noting that at least one had used an AI chatbot to write a report for them.

In a 223-page injunction ruling Thursday, U.S. District Judge Sara Ellis cited body-worn camera footage to document the many ways federal agents made unreliable and false statements when describing their several violent clashes with Chicago residents, the Chicago Tribune reported.

“The Court also notes that, in at least one instance, an agent asked ChatGPT to compile a narrative for a report based off of a brief sentence about the encounter and several images,” Ellis wrote in one footnote.

“To the extent that agents use ChatGPT to create their use of force reports, this further undermines their credibility and may explain the inaccuracy of these reports when viewed in light of the BWC footage,” she added.

This is particularly concerning because, as OpenAI’s own website states: “ChatGPT can be helpful—but it’s not always right.”

Ellis also accused Border Protection Commander Gregory Bovino of giving testimony that was specifically “not credible,” either responding to questions with “cute” answers or “outright lying.”

An appeals panel blocked Ellis’s injunction Wednesday, stating that it was “overbroad” in its application to the government, but asserted that her findings could “support entry of a more tailored and appropriate preliminary injunction.”

Ellis’s injunction would have barred federal agents from using tear gas and other riot prevention methods against protesters, “unless such force is objectively necessary to stop an immediate threat.”

It seems that using ChatGPT may be an emerging, and concerning, trend in the armed forces. Last month, a U.S. Army general told reporters that he’d become very fond of “Chat,” even trusting the algorithm to make “key command decisions” in relation to his post.