Anthropic Draft Gives US Government AI Release Power

Anthropic presented US lawmakers with a draft bill that would let the government block an artificial intelligence model’s release under specified conditions, putting a potential veto over deployment in public hands. The proposal covers models judged to pose an imminent risk of killing or injuring more than 50 people or causing more than $1 billion in damages. The article identifies no demographic or cultural effects from the draft.

Who Would Decide

Under Anthropic’s proposal, the government would decide whether a model presents the specified imminent risk and could then block its release. The threshold concerns projected harm: more than 50 people killed or injured, or more than $1 billion in damages. The draft thus ties government intervention to a defined scale of potential consequences, rather than setting out a general prohibition on AI releases.

The proposal names three ways the risk could arise: a model creating or releasing a biological weapon, an AI going beyond human control, or a cyberattack. The article doesn’t provide the draft’s full text, name the government body that would make the determination, or describe how someone could challenge a block. It also doesn’t say whether lawmakers have accepted or acted on the proposal.

Politico first reported the proposal. Anthropic presented it to US lawmakers amid heightened concern about risks from rapidly advancing technology. The article names no other institutional backers or beneficiaries and doesn’t describe a broader international framework for the measure.

Threat Reporting and Legal Limits

Separately, legal experts questioned whether existing US law requires AI companies to report threats of violence. Current duties to warn can apply to mental health professionals or parties with a special relationship, but the experts said those duties may not extend to AI companies. The account leaves that question unresolved.

The issue has been raised by the February 10, 2026, mass shooting in Tumbler Ridge, British Columbia, and lawsuits alleging OpenAI failed to prevent it. Those lawsuits are allegations, not findings described in the article. The article doesn’t say the draft bill was prompted by the shooting or that it would require AI companies to report threats.

Privacy, Monitoring and Risk Assessment

Legal experts also raised concerns about privacy, false threat reports and unequal monitoring. They discussed how difficult it can be to distinguish a credible threat from fictional writing, research or intrusive thoughts. Those distinctions matter to the separate question of whether companies should report threats; the article doesn’t say the draft bill addresses them.

The source describes no nationalist or dissident response, no public protest, and no evidence of demographic replacement or cultural dispossession connected to the proposal. Its central policy question is narrower: should government be able to stop a model’s release when officials determine it could cause mass casualties or more than $1 billion in damage, and does existing law require AI companies to report violent threats? The proposal would move a specific release decision toward government authority, while legal experts’ questions leave companies’ warning duties unsettled.