Science & Tech · AI · 3 hrs ago
Anthropic proposes bill to pause AI models posing catastrophic risks
Anthropic is circulating a draft AI safety bill for Congress that could stop some models from being released.
It defines an imminent catastrophic risk as one that could put more than 50 people at risk of death or cause over $1 billion in property damage.
The danger must stem from a biological weapon, a model evading its developer’s controls, or a cyberattack.
Independent evaluators registered with the Commerce Department would assess whether a model poses that risk, and could be paid by AI companies as long as payment did not depend on the findings.
If evaluators find an imminent risk, they would have 72 hours to notify the department, which would then have to ask a Washington, DC, judge to halt the release.
Federal agencies would not be allowed to stop distribution on their own.
The proposal has not been described as enacted, and it is unclear how the system would ensure that company payments do not influence evaluations.
Anthropic is proposing a draft bill that could pause the release of AI models posing an imminent catastrophic risk.
The draft covers risks involving more than 50 deaths or more than $1 billion in property damage.
It ties those risks to biological weapons, a model evading its developer’s controls or a cyberattack.
Independent evaluators would assess models, and the U.S. Commerce Department would ask a D.C. judge to halt a release.
The proposal would not let a federal agency halt distribution on its own.
- Who
- Anthropic proposed the draft bill.
- What
- The bill would allow a judge to halt an AI model’s release if it poses a foreseeable, imminent catastrophic risk.
- When
- TechCrunch published its report on October 20, 2026. Anthropic said it had advocated for the framework since June.
- Where
- The proposal concerns U.S. federal regulation, with a D.C. judge involved in any request to halt a release.
- Why
- The bill aims to address risks involving mass deaths or major property damage linked to biological weapons, a model evading developer controls or a cyberattack.
This story does not have two clearly opposing sides.
Anthropic said it had advocated for proposals in its Advanced AI Framework with federal and state lawmakers.
TechCrunch reported that Anthropic was circulating a draft AI safety bill for Congress to consider.
- Death threshold
- More than 50 people at risk of dying
- Property damage threshold
- More than $1 billion
- Evaluator notification deadline
- 72 hours
- Evaluator registration
- U.S. Commerce Department
- Judicial request
- The Commerce Department would ask a D.C. judge to halt a model’s release
What people are saying
This bill represents the proposals outlined in Anthropic’s Advanced AI Framework, which we have been advocating for with federal and state lawmakers since June