Anthropic rejects Pentagon requests in AI safeguards dispute, CEO says

Anthropic is refusing to remove restrictions on its AI system Claude despite pressure from the Pentagon, escalating a dispute over how artificial intelligence can be used within the US military.

The standoff centers on safeguards Anthropic has placed on Claude — the first AI system deployed inside the military’s classified network — that prevent it from being used for fully autonomous weapons targeting or mass domestic surveillance.

Anthropic CEO Dario Amodei said Thursday the company cannot agree to the Defense Department’s request to loosen those restrictions, even after officials warned the firm could be removed from government systems.

“Regardless, these threats do not change our position: we cannot in good conscience accede to their request,” Amodei said.

According to Amodei, the department has indicated it will contract only with AI companies that agree to “any lawful use” of their systems, a condition that would require removing the safeguards Anthropic insists on maintaining.

Anthropic, backed by Google and Amazon, holds a contract with the Defense Department worth up to $200 million.

– Ends