Anthropic shops draft bill that could block risky AI models
Original title: Anthropic pitches AI safety bill that may block a model if 50+ people could be killed
Evaluators would have 72 hours to notify Commerce, which must ask a D.C. judge before a release can be halted.
Politico reported that Anthropic is circulating a draft AI safety bill it wants Congress to consider. A model could be blocked for a foreseeable, imminent risk of more than 50 deaths or over $1 billion in property damage from a biological weapon, the model evading developer controls, or a cyberattack. Independent evaluators registered with the Commerce Department, possibly paid by labs if payment is not tied to results, would flag risk. An Anthropic spokesperson confirmed the draft and tied it to the company’s Advanced AI Framework, advocated since June.
Unconfirmed: The text says it is unclear how the draft would stop lab payments from influencing evaluations, and Congress has not adopted the bill.
- Thresholds mirror California Senate Bill 53’s definition of catastrophic risk.
- No federal agency could halt distribution on its own discretion.
- Anthropic confirmed the draft to Politico but did not immediately answer TechCrunch.
- Nvidia CEO Jensen Huang has argued the industry needs no new laws.
- Organizations
- AnthropicU.S. Commerce DepartmentDeepMindNvidia
- Countries
- United States
From the original article
3 min read at techcrunch.com
The proposed policy seems to offer protections against AI run amok but actually forbids federal agencies from blocking a model's release without the permission of a judge.
By Lucas Ropek · techcrunch.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.