Anthropic deepened a high-stakes standoff with the Pentagon as CEO Dario Amodei said the company would not grant the Defense Department unrestricted access to its Claude AI, citing concerns over mass surveillance of Americans and fully autonomous weapons. Defense Secretary Pete Hegseth set a Friday deadline for Anthropic to allow “all lawful uses” or face cancellation of its government contract, with officials also weighing a supply‑chain‑risk designation or use of the Defense Production Act. Anthropic said it remains in talks but will help transition the military to another provider if needed. The Pentagon countered that it has no interest in illegal surveillance or weapons operating without human involvement. The clash underscores tensions between rapid military adoption of generative AI and corporate safety policies, even as peers such as Google, OpenAI and xAI supply models to a new internal defense network. Lawmakers from both parties criticized the Pentagon’s handling and urged stronger AI governance for national security.
Related article:
