The Pentagon previously added Anthropic, the maker of the Claude AI chatbot, to a list of companies deemed supply chain risks, effectively barring it from new War Department contracts. President Donald Trump soon after added that federal agencies would “immediately cease” the use of Anthropic systems. 

Warren questioned whether the designation was based on legitimate national security concerns or punishment for the company’s stance on AI ethics. She asked the department to provide documentation supporting its decision and explain whether political considerations played a role. 

The move comes amid a broader dispute between the government and Anthropic over the use of AI models, including uses tied to autonomous weapons and mass surveillance, with Anthropic citing safety and ethical concerns. The Pentagon disputed the allegations. 

Anthropic has moved forward with a lawsuit against the Pentagon over its blacklisting, arguing the decision is unlawful and could significantly harm its business. The company has said it is being penalized for maintaining safeguards on its technology rather than for any legitimate security issues. 

In response, the Pentagon filed a rebuttal lawsuit in which it argued that Anthropic’s prior access to the department’s “technical and operational warfighting infrastructure would introduce unacceptable risk into DoW supply chains.” It asked the U.S. District Court for the Northern District of California to bar Anthropic from moving forward with its lawsuit. 

A preliminary hearing for Anthropic’s initial suit is scheduled for Tuesday in California.

Days after the Pentagon cut ties with Anthropic, it announced a deal with OpenAI, which said the use of its AI systems for mass surveillance or fully autonomous weapons would not be possible because of its “safety stack,” existing laws, and the contract language, CNBC reported, though the contract has not been shared in full.

Warren also asked for the terms of OpenAI’s agreement with the department in her letter. 

ORGANIZATIONS WITH STRONG TIES TO THE LEFT INFLUENCED AI POLICY IN A DEEP RED STATE

Elon Musk’s xAI signed a contract with the Pentagon last month, allowing the military to use the company’s Grok large language model. Claude was previously the only model used by the military for sensitive intelligence and combat operations.

The Washington Examiner reached out to Warren’s office for comment.