Anthropic pitches AI safety bill to U.S. lawmakers

Traders Union
All News
Regulators
Anthropic AI safety bill

Highlights

Anthropic's draft bill proposes government intervention only if an AI model poses imminent risks exceeding $1 billion in damages or threatens over 50 lives.
Independent evaluators registered with the Commerce Department would periodically assess frontier AI models, with findings required within 72 hours for imminent risks.
The draft limits executive action, allowing suspensions of AI models solely via orders from a federal judge in the District Court for the District of Columbia.

Draft proposal sets limits for intervention

As first reported by Reuters, citing Politico, Anthropic representatives have been discussing a proposed bill with Capitol Hill lawmakers that would regulate frontier AI safety and security. The draft would allow the government to block a model's release or use only if it creates a foreseeable, imminent risk of more than $1 billion in damages, or of killing or seriously injuring more than 50 people.

The reported thresholds apply when the risk stems from the creation or release of a biological weapon, a model evading its developer's controls, or a cyberattack. Anthropic and the Department of Commerce did not immediately respond to Reuters requests for comment.

Under the draft, independent evaluators registered with the Commerce Department would periodically assess AI labs and their frontier models. Those labs could pay the evaluators, provided the payments are not tied to the outcome of the review.

Commerce role and industry implications

If evaluators find what the draft describes as an imminent enumerated risk in a company's frontier models, they would have to notify the Commerce Department within 72 hours of completing their assessment. The department could then ask a judge to halt a model after determining that the risk meets the bill's standard.

The language also reportedly limits broader executive action by stating that no other U.S. government department, agency or officer could suspend a model's development, deployment or internal use. Any suspension order would instead come from a federal judge at the District Court for the District of Columbia, outlining a court-centered process for intervention in the AI sector.

In our earlier article, we covered the intensifying debate over how governments and major developers should manage AI risks, including calls for deeper international coordination. We noted proposals for U.S.-China channels to handle serious AI incidents and discussed the idea of the UK serving as a neutral venue for broader AI safety dialogue as cross-border threats grow.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.