Judge blasts immigration agents using ChatGPT to write ‘inaccurate’ reports
Agents criticised for using ChatGPT for immigration reports
- A federal judge criticised immigration agents for using artificial intelligence, specifically ChatGPT, to draft use-of-force reports, citing concerns about accuracy and credibility.
- U.S. District Judge Sara Ellis noted factual discrepancies between official law enforcement narratives and body camera footage, observing an agent instructing ChatGPT to compile a report from a brief description and images.
- Experts condemned this practice as the “worst possible use” of AI, highlighting significant risks to accuracy, as AI-generated reports may not reflect the officer's actual experience or perspective.
- The Department of Homeland Security did not comment on its policies regarding AI use by agents, and experts noted that few departments have established clear guidelines for this technology.
- Concerns were also raised about privacy, particularly if public versions of ChatGPT are used, and the inherent difficulties in using AI with visual components for accurate report generation.