Anthropic AI Safety Bill Could Let US Halt Models Posing Catastrophic Risks

Anthropic's proposed US legislation to suspend AI models that pose catastrophic risks.

The United States could gain a legal mechanism to suspend artificial intelligence models that pose catastrophic risks under a proposed bill developed by Anthropic, according to information reported by Politico and shared by Coin Bureau. The draft would allow the government to halt an AI model if it presents a “foreseeable, imminent” threat capable of causing $1 billion in damage or resulting in more than 50 deaths and injuries.

The proposal outlines a potential framework for government intervention when an AI system is assessed as posing an immediate and severe threat. The risks identified include the development of biological weapons, cyberattacks and situations in which an AI model escapes its developer's control.

Despite the potential scope of the proposed authority, the draft would impose specific restrictions on which government bodies could initiate the process. The bill has not been formally introduced, meaning the provisions remain part of a proposal rather than enacted law.

Commerce Department Would Hold Exclusive Authority

Under the draft, only the US Commerce Department would be permitted to ask a federal judge to suspend an AI model over the specified risks. Other federal agencies would be barred from independently seeking such a suspension under the proposed framework.

This provision would centralize the authority to initiate judicial intervention within a single department. Rather than allowing multiple agencies to independently pursue the suspension of an AI model, the proposal would establish the Commerce Department as the sole federal body authorized to request the measure from a judge.

The requirement to approach a federal judge is another central element of the draft. The proposed mechanism would therefore involve a judicial process rather than granting every federal agency independent authority to suspend a model under the framework described in the report.

Independent Evaluators Would Have 72 Hours

The proposal would also establish a specific timeline for assessing potential threats. Independent evaluators would have 72 hours to report risks considered imminent, creating a defined reporting period for identifying potentially catastrophic dangers associated with an AI model.

The 72-hour requirement is a key operational detail in the draft, although the available information does not specify the full evaluation procedures, the criteria for selecting evaluators or the precise steps that would follow a report identifying an imminent threat.

The proposal's stated risk threshold combines potential financial damage and human harm. It identifies scenarios involving $1 billion in damage or more than 50 deaths and injuries, alongside the requirement that the danger be “foreseeable, imminent.” The draft specifically references bioweapons, cyberattacks and AI systems escaping their developers' control as examples of the types of threats under consideration.

Proposal Has Not Been Formally Introduced

According to Coin Bureau, citing Politico's reporting, the draft was proposed by Anthropic as a framework for addressing catastrophic AI risks. Its provisions could give the US government a defined route to seek the suspension of a dangerous model while limiting which agencies can initiate that process.

However, the bill has not been formally introduced. Its current status means the proposed authority, reporting timeline and restrictions on federal agencies should not be treated as existing legal requirements. The draft's eventual wording, legislative progress and any changes to its provisions remain undetermined based on the information reported.

Writer: Victoria Hale
Technology & Blockchain Writer

Victoria Hale writes about blockchain technology, digital infrastructure, and the intersection of emerging technologies with finance. Her articles explore how new protocols and systems are shaping the evolving digital economy.

She prioritises clarity and accuracy when explaining technical developments to a general audience.

Check out other news and articles on Google News

Disclaimer:

The articles on Hokanews are here to keep you updated on the latest buzz in crypto, tech, and beyond—but they’re not financial advice. We’re sharing info, trends, and insights, not telling you to buy, sell, or invest. Always do your own homework before making any money moves.
Hokanews isn’t responsible for any losses, gains, or chaos that might happen if you act on what you read here. Investment decisions should come from your own research—and, ideally, guidance from a qualified financial advisor. Remember: crypto and tech move fast, info changes in a blink, and while we aim for accuracy, we can’t promise it’s 100% complete or up-to-date.