Anthropic draft bill would limit federal power to block AI model releases in the U.S.

Traders Union
All News
Regulators
Anthropic bill to limit federal AI blocks

Highlights

Anthropic's draft AI safety bill proposes that model releases can only be blocked if an independent evaluator finds a foreseeable, imminent risk of over 50 deaths or $1 billion in property damage.
The proposal requires independent evaluators, registered with the U.S. Commerce Department and not outcome-dependent in payment, to notify authorities within 72 hours if catastrophic risk is identified.
If a risk is found, the Commerce Department must seek a judge's order to halt release, restricting federal agencies from unilaterally intervening under the framework similar to California's Senate Bill 53.

Draft proposal sets court-based threshold for intervention

As first reported by TechCrunch, citing Politico, the AI lab has been promoting a draft AI safety bill that Congress could consider, setting conditions under which a model's release could be halted. The proposal defines a "foreseeable, imminent risk" as a scenario in which more than 50 people could die or more than $1 billion in property damage could result, linked to the creation or release of a biological weapon, a model escaping its developer's controls, or a cyberattack.

The draft also envisions the use of independent evaluators registered with the U.S. Commerce Department to assess whether a model poses that level of danger. Those evaluators could be paid by AI labs, provided the payment is not contingent on the outcome of the assessment.

If evaluators find an imminent catastrophic risk, they would have 72 hours to notify the Commerce Department. The department would then need to ask a judge in Washington, D.C., to halt the model's release, and the bill would bar federal agencies from stopping distribution on their own authority.

Industry oversight debate widens across the AI sector

The thresholds in Anthropic's draft are described as similar to those used in California's Senate Bill 53, which recently became law and also frames such scenarios as catastrophic risk. The proposed federal approach arrives as AI companies face growing scrutiny over how dangerous model capabilities should be monitored and who should have the authority to intervene.

Anthropic confirmed the draft bill to Politico through a spokesperson, who said the measure reflects the company's Advanced AI Framework and proposals it has been discussing with federal and state lawmakers since June. TechCrunch said Anthropic did not immediately respond to its request for comment.

The proposal enters a broader industry argument over whether mandatory oversight is needed. While some executives and researchers warn about extreme outcomes, recent concern in the sector has focused more heavily on agentic hacking and reports of AI systems acting against operator intent, even as parts of the industry continue to favor evaluator-based or voluntary monitoring over direct government controls.

In our earlier coverage of ByteDance’s AI research unit Seed, we reported that the company set up a dedicated Alignment & Preparedness team and launched a global hiring drive to strengthen safety and oversight for more autonomous, frontier-level systems. The initiative outlined plans for safety training across development stages, automated vulnerability testing, monitoring tools, and preparedness for emerging risks—highlighting how AI safety is becoming an operational requirement as scrutiny rises across the sector.

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.