Trump Bans Anthropic AI From All U.S. Agencies After Safety Fight

For months, Anthropic and the Pentagon held private talks over AI access. Anthropic wanted clear limits: no mass surveillance of Americans and no autonomous weapons. The Pentagon demanded full, unrestricted access, leading to a deadlock.

CEO Dario Amodei publicly refused to give in, saying Anthropic “cannot in good conscience accede.” He defended 2 firm red lines: no domestic surveillance and no fully autonomous weapons, saying these principles could not be compromised.

President Trump ordered all U.S. agencies to stop using Anthropic’s AI, calling it “radical left” and “woke.” Hegseth labeled it a supply chain risk, and Trump warned of civil and criminal consequences during a 6-month phase-out.

On Friday, Anthropic called the action “unprecedented and legally unsound,” saying such a label was never used on a U.S. company. It promised to fight the supply chain risk decision in court and refused to change its safety stance.

Hours later, Sam Altman secured a Pentagon deal for classified networks but kept the same red lines: no mass surveillance and human control of weapons. Elon Musk backed Trump, and his chatbot Grok also gained Pentagon access.

The clash shook Silicon Valley. Investors and AI workers supported Anthropic. Gen. Shanahan warned “everyone loses,” calling the red lines reasonable. Senator Warner questioned political influence, as the AI warfare debate remained unresolved.





