“Wildly irresponsible”: DOT’s use of AI to draft safety rules sparks concerns (arstechnica.com)

The U.S. Department of Transportation (DOT) is reportedly using artificial intelligence, specifically Google Gemini, to draft safety regulations for transportation, raising significant concerns about accuracy and safety. This approach, revealed in a ProPublica investigation, is alarming experts who fear that the technology's tendency to generate false or "hallucinated" information could lead to flawed legislation with dire consequences. Staff members worry that errors in AI-generated rules could lead to lawsuits, injuries, or even fatalities within the transportation sector. Despite these concerns, DOT's top lawyer, Gregory Zerzan, is primarily focused on efficiency rather than perfection, asserting that the goal is to expedite the rulemaking process from weeks to under 30 days. He emphasized a "good enough" approach, highlighting that Gemini can produce draft regulations in under half an hour. While some experts believe AI could serve as a valuable research assistant with proper oversight, many within the DOT remain skeptical about relying on it for creating critical safety rules, underscoring a tension between innovation and the responsibility to public safety in the application of AI in governance.