US Senate negotiators are drafting legislation to regulate frontier artificial intelligence developers. The proposed bill mandates that companies implement safety measures at the design stage to mitigate catastrophic risks, such as the potential use of AI in developing nuclear or biological weapons.
The bill introduces a duty of care for developers of advanced models, specifically targeting major industry players like OpenAI, Google, and Anthropic. It grants the federal government authority to halt the release of models deemed to pose an imminent threat, while allowing companies to challenge such decisions in federal court.
Legislators aim to establish a unified federal standard for AI safety, prohibiting individual states from enacting conflicting regulations. This approach seeks to prevent a fragmented regulatory landscape across the country.
Key figures involved in the bipartisan negotiations include Senators John Thune, Ted Cruz, Amy Klobuchar, and Maria Cantwell. The group is currently refining the specific criteria and triggers for federal intervention.
The legislative timeline remains tight due to the upcoming US midterm elections. With only a few weeks of session remaining, the bill faces significant procedural hurdles to pass both chambers of Congress before the recess.
