Leaked AI safety bill from Anthropic only requires government intervention if 50+ injuries or $1 billion are at risk

A leaked draft US bill reportedly backed by Anthropic would limit when the federal government could intervene over risks from advanced AI systems. It sets high thresholds for action, raising questions about whether the proposed safeguards would address serious harms that fall below them.

Under the draft, intervention would be required only when a system poses a risk of at least 50 injuries or $1 billion in damage. The supplied article excerpt gives no further details about the bill’s provisions, its status, or Anthropic’s response.