President Trump bans use of AI from federal agencies after Anthropic refuses to unlock capabilities — Anthropic raises risks of autonomous military use, domestic mass surveillance

President Donald Trump ordered all U.S. federal agencies to stop using technology from AI company Anthropic on Friday, February 27, posting the directive on Truth Social at 3:47 p.m. ET. This was more than an hour before the Pentagon’s deadline of 5:01 p.m. ET for Anthropic to comply with the request.

“America will never allow a radical left-wing company to dictate how our great military fights and wins wars,” Trump said, furious at Truth Social, adding that he was directing all federal agencies to “immediately stop using Anthropic’s technology.”

The dispute stems from a contract Anthropic signed with the Department of Defense last summer worth up to $200 million. Anthropic had asked for written assurances that its Claude models would not be used for domestic mass surveillance of U.S. citizens or to control weapons systems capable of firing without human involvement. The Pentagon countered that it needed the right to deploy the Claude for “any lawful purpose,” and argued that it would be impossible to negotiate exemptions for individual use cases with private companies.

Mr. Amodei said Thursday that his company “cannot in good conscience” comply with the Pentagon’s terms, after months of private talks broke down in a public standoff this week. In response, the Pentagon threatened to invoke the Korean War-era Defense Production Act to force Anthropic into compliance and warned that it would designate Anthropic as a “supply chain risk,” a label typically given to companies from hostile countries such as Huawei.

In a post on Truth Social, President Trump accused Anthropic of trying to “strong-arm the Department of the Army and force it to follow the Terms of Service instead of the Constitution,” adding that the company’s position “puts American lives at risk.” He gave the agencies a six-month phase-out period and warned that if Anthropic fails to cooperate during that time, he will “use the full power of the President to compel the government to comply, with serious civil and criminal consequences.”

Claude is the only AI model approved for use in classified military systems, and defense software company Palantir, which uses Claude for its most sensitive government contracts, urgently needs to find a replacement. OpenAI CEO Sam Altman said Friday that he shares Anthropic’s position on the ethical “red line” of autonomous weapons, complicating his candidacy as a direct successor.

Google's preferred source