Bipartisan Senate negotiators John Thune, Ted Cruz, and Amy Klobuchar are debating legislation to impose a duty of care on artificial intelligence companies. The legal standard requires developers of advanced models to prevent catastrophic risks before product release. The bill specifically targets firms developing high-level AI, including Google, OpenAI, and Anthropic.
The proposal grants the U.S. government authority to block AI models deemed unsafe, particularly those capable of designing nuclear or biological weapons. Affected companies would maintain the right to challenge government intervention in federal court.
The legislation aims to create a unified federal framework by pre-empting states from enforcing independent AI risk laws. Congressional leaders consider passage before the November midterm elections uncertain due to a limited legislative calendar.