The dispute escalates: The US Department of Justice is demanding that Anthropic be removed from all federal agencies, claiming it could shut down AI during wartime.

According to 1M AI News, the U.S. Department of Justice filed documents with the court on Tuesday requesting the expulsion of Anthropic from all federal agencies, expanding the scope of the dispute from a previous direct contract with the Department of War to the entire executive branch. The documents state, "For national security reasons, Anthropic's AI technology service terms are no longer acceptable to the executive branch." The Justice Department disclosed previously undisclosed details of the negotiations in the documents: In early 2026, the Department of War requested a clause in the contract allowing the use of Anthropic technology for "any lawful purpose," which Anthropic rejected, citing its Claude usage policy. Emil Michael, Under Secretary of Defense Research and Engineering, stated in an affidavit that Anthropic displayed "hostility" during the negotiations, and its position "appeared primarily to serve its public image rather than be based on facts." The most controversial section of the document warns that "AI systems are highly susceptible to manipulation, and if Anthropic determines that its corporate red lines have been crossed, it may attempt to shut down its technology or prematurely alter the behavior of its models during ongoing wartime operations." Anthropic had previously requested a commitment from the U.S. that its AI would not be used for mass surveillance of U.S. citizens or for the deployment of autonomous weapons. An Anthropic spokesperson responded that it was reviewing government documents and reiterated that "seeking judicial review does not change our long-standing commitment to using AI to protect national security." The hearing is scheduled for next Tuesday.