According to a report by ProPublica, the U.S. Department of Transportation (DoT) is planning to use Google's Gemini AI to assist in drafting binding regulations related to aviation, automotive, rail, and maritime safety. This move aims to significantly shorten the policy-making cycle and achieve "lightning-fast" release of regulatory rules.
Image source note: The image was generated by AI, and the image service provider is Midjourney.
In internal communications, DoT lawyer Daniel Cohen demonstrated the "revolutionary" efficiency of AI to employees. While traditional regulation drafting usually takes months or even years, the demonstration showed that Gemini could generate drafts in "a few seconds." DoT's General Counsel, Gregory Zerzan, openly stated, "We don't need perfect rules in the XYZ field, nor even very good rules. We aim for 'good enough,' with the goal of 'covering all bases.'"
However, this "speed-first" strategy has raised strong concerns among industry experts:
Risk of "hallucinations": Gemini has previously had multiple instances of fabricating information or medical errors; applying it to safety regulations where even small mistakes can have major consequences (such as air traffic control) poses obvious potential risks.
Professional doubts: Mike Horton, former Chief AI Officer at DoT, compared this plan to "hiring a high school intern to draft regulations."
Accountability issues: Zerzan pointed out that Trump was "very excited" about this plan, aiming to reduce drafting time to within 30 days. However, legal experts worry that rules lacking thorough review may lead to compliance chaos.
