Judge Blocks Pentagon's Attempt to Blacklist Anthropic

A federal judge has overturned a Department of Defense decision to designate AI lab Anthropic as a national security supply-chain risk, ruling the move was "illegal and baseless." As Wired first reported, U.S. District Judge Rita Lin determined that the designation, issued by Defense Secretary Pete Hegseth on February 27, constituted unconstitutional retaliation.

The ruling vacates the blacklist that made Anthropic ineligible for federal contracts and lifts punitive measures that barred military suppliers from doing business with the company. Judge Lin described those measures as "arbitrary, capricious, an abuse of discretion, and otherwise not in accordance with law." The decision also removes sanctions improperly imposed on Anthropic by nine agencies, including the Treasury Department, the State Department, the Pentagon, and the Department of Homeland Security.

Wired reports the conflict began during negotiations over a $200 million deal to utilize Anthropic's Claude models for military use. The dispute intensified following reports that Claude was employed by the U.S. in an operation to capture Venezuelan president Nicolás Maduro. Anthropic sought limits on the deployment of its models regarding mass surveillance systems and lethal autonomous weapons, which Hegseth rejected, arguing that contractors cannot dictate the use of technology once deployed.

While the Pentagon argued that Anthropic's potential to alter or disable technology created an "unacceptable risk," Judge Lin noted that the government continues to discuss collaborating with Anthropic on its new model, Mythos, in sensitive contexts. While the ruling restores Anthropic's eligibility for contracts, Lin affirmed that the Pentagon is not required to use the company's models. A Pentagon spokesperson was unavailable for comment, though the department is expected to appeal.