On March 18, Bloomberg reported that the Trump administration vowed to use legal means to remove Anthropic’s AI technology from all U.S. government agencies. Previously, the two parties had disputes over restrictions on the use of Anthropic AI technology.
Anthropic feuds with US military
"Plaintiff Anthropic's terms of service for its AI technology have been deemed unacceptable by the Executive Branch for national security reasons," the U.S. Department of Justice, representing the government, said in a court filing on Tuesday.
Anthropic developed the AI chatbot Claude. The company has filed a lawsuit in an attempt to prevent the Department of Defense from labeling it a risk to U.S. supply chains, escalating a dispute over restrictions on the military's use of AI. Anthropic wants restrictions on how its products can be deployed, arguing that a protracted legal battle could cost it billions of dollars in lost revenue. Anthropic is asking the court for a preliminary injunction to prevent the government's ban from taking effect while the lawsuit proceeds.
Lawyers for the U.S. Department of Justice stated in court documents submitted that in early 2026, the U.S. Department of Defense negotiated with Anthropic on the contract between the two parties, hoping to add a clause that would allow the Pentagon to use its technology for any legal purpose, but Anthropic "refused" to accept the clause because of its Claude use policy.
The U.S. Department of Justice said that throughout the negotiation process, "Anthropic's overall performance caused the Department of Defense to question whether Anthropic was still a trustworthy partner with whom it was willing to contract in this highly sensitive area."
The U.S. Department of Justice is also concerned that allowing Anthropic to continue to have access to its technology and combat systems "will create unacceptable risks to the Department of Defense supply chain."
"After all, AI systems are extremely susceptible to manipulation. If Anthropic determines on its own that its company's 'red line' has been crossed, it may attempt to disable its technology or preemptively change the behavior of its models before or during combat operations," a U.S. Department of Justice lawyer said.
At the direction of U.S. President Donald Trump, the Pentagon and other federal agencies are shifting their AI operations to other vendors, based on a risk assessment typically reserved for foreign companies. Anthropic had asked the government for assurances that its AI would not be used to conduct mass surveillance of Americans or deploy autonomous weapons.
U.S. Defense Secretary Hegseth said in a social media post on February 27 that the U.S. military will continue to use Anthropic’s technology tools “but not for more than six months to allow for a smooth transition to a better, more patriotic service provider.”