Stories
Anthropic AI Safety Threshold Sparks Regulatory Debate
F
Firstpost•19-10-2026, 02:17
Anthropic AI Safety Threshold Sparks Regulatory Debate
- •Anthropic proposed draft legislation allowing the government to block AI models posing severe risks to the public.
- •The proposed threshold for intervention requires a foreseeable risk of more than 1 billion dollars in damages or killing and seriously injuring more than 50 people.
- •Intervention would be strictly limited to three categories including biological weapons, AI evading controls, and cyberattacks.
- •The bill seeks to centralize regulatory authority within the Department of Commerce and could limit states from introducing stricter rules.
- •Anthropic has significantly increased its lobbying efforts in Congress and at the state level to shape upcoming AI regulation.
Key Questions Answered
Under the draft bill proposed by Anthropic, the government would only be permitted to block a model's release or use if it presents a foreseeable and imminent risk of causing more than $1 billion in damages or resulting in the death or serious injury of more than 50 people.
The proposed legislation restricts government intervention to three distinct categories of harm: the creation of a biological weapon, an AI model evading the controls set by its developer, or the execution of a cyberattack.
The bill specifies that no US agency other than the Department of Commerce would be permitted to intervene to suspend the development or internal use of an AI model.
