What We Know
Anthropic has quietly proposed or circulated a draft federal AI safety bill that would give the government a way to seek court intervention against an AI model.2Backed by 2 sourcesppc.landeffectstory.com Under the proposal, courts could halt an AI model for up to 45 days, according to the draft’s reported provisions.1Backed by 1 sourcesppc.land
The bill’s intervention threshold is described as a risk that an AI model could contribute to at least 50 deaths or cause $1 billion in damages.3Backed by 3 sourcesTechCruncheffectstory.comtomshardware.com The draft would center the Commerce Department in seeking a court halt when those thresholds are met, rather than giving another agency direct authority to stop the model.2Backed by 2 sourceseffectstory.comtomshardware.com
The proposed threshold reportedly mirrors a California law while being repurposed to determine when the government could block an AI model.1Context from one sourcePolitico The draft was leaked and differed from positions Anthropic CEO Dario Amodei had advocated earlier.1Context from one sourceThe Washington Post Anthropic quietly pitched its own AI safety bill to lawmakers, according to reporting attributed to Politico and Reuters.1Context from one sourceinvesting.com
Source Comparison
Aligned reportingCorroborates
- ppc.land↗Its headline directly supports the reported Anthropic-linked draft and the proposed 45-day court halt.
- TechCrunch↗Its headline supports the reported threshold involving a potential death toll of at least 50 people.
- effectstory.com↗Its headline supports the proposed government court-halt mechanism, the 50-life or $1 billion threshold, and Commerce Department responsibility.
- tomshardware.com↗Its headline supports both the proposed 50-death or $1 billion trigger and the Commerce Department’s central role in stopping an AI model.
Adds context
- Politico↗It adds that the proposed $1 billion threshold mirrors a California law but would be repurposed to determine when the government could block an AI model.
- The Washington Post↗It contributes that the federal AI legislation draft was leaked and differed from positions Anthropic CEO Dario Amodei had advocated earlier.
- investing.com↗It adds that Anthropic quietly pitched its own safety bill to lawmakers, with the report attributed to Reuters and Politico.