U.S. Senate negotiators are considering legislation requiring AI developers to show they are taking reasonable precautions against harm. The proposal would give the U.S. commerce secretary authority to demand evidence of those safeguards and deploy government auditors to test products. The meaning of "reasonable" has not been determined, and the legislation remains under debate with long odds of becoming law. U.S. President Donald Trump said existing authorities already provide adequate tools to regulate and prosecute technology companies, suggesting he would not sign new AI rules. The talks involve Senate Majority Leader John Thune, Senate Commerce Committee Chairman Ted Cruz and Senator Amy Klobuchar, the top Democrat involved. The debate follows warnings from Anthropic researcher Jacob Coxon, who said he resigned partly because some AI builders believe the technology could kill everyone by the end of the decade, as well as a worldwide decline in AI-related stocks after industry leaders urged a slower development pace.