Anthropic Reaffirms Refusal to Remove AI Safeguards Amid Pentagon Pressure

Dario Amodei, CEO of Anthropic, reaffirmed that the company will not remove safeguards restricting certain uses of its AI systems, even under pressure from the Pentagon. The statement follows earlier reports of Pentagon demands for unrestricted access to Anthropic’s Claude model, according to a statement published on Anthropic’s website.

Amodei said Anthropic will not allow its AI to be used for mass domestic surveillance or in fully autonomous weapons systems. He stated that such uses are either incompatible with democratic values or beyond the safe capabilities of current AI technology. The company emphasized that it supports lawful foreign intelligence operations and partially autonomous defense systems but will not provide products that remove human oversight in lethal decision-making.

According to the statement, the U.S. Department of War has threatened to designate Anthropic a “supply chain risk” or invoke the Defense Production Act to compel compliance. Amodei called these actions contradictory, noting that one implies the company is a security threat while the other deems its technology essential to national defense.

Anthropic indicated it remains willing to collaborate with the Department on research to improve AI reliability under proper oversight. The company stated that if offboarded, it would assist in a smooth transition to another provider to prevent disruption to military operations.