A U.S. federal judge on Thursday temporarily halted the Pentagon's attempt to place Anthropic on a national security blacklist that would limit the company's eligibility for government contracts. U.S. District Judge Rita Lin ruled the government may not immediately implement its designation of Anthropic as a national security "supply-chain risk."
The court's order prevents the label - which, if enforced, would restrict Anthropic's participation in federal contracting - from taking immediate effect while the matter proceeds through the courts. The judge indicated in her ruling that the administration's actions could have been motivated by a desire to punish the company for its public positions on AI safety rather than by a strictly national security-driven rationale.
Anthropic had challenged the designation on constitutional grounds, arguing it infringed on the company's free speech rights and deprived it of due process. The company also said it was not afforded an opportunity to contest the decision before the label was applied.
The Pentagon has defended the decision to apply the supply-chain risk designation, describing the step as lawful and necessary to protect military systems from potential risks. The agency framed the move as a protective measure intended to shield defense technology and operations.
The underlying conflict centers on Anthropic's policy decisions regarding the uses of its AI models. Anthropic has refused to permit its systems to be used for autonomous weapons applications or for domestic surveillance, a stance that has been cited as a root cause of the dispute with the Pentagon.
Judge Lin's ruling will be delayed briefly to allow the government an opportunity to appeal, meaning the court's temporary block is not yet final. While the decision represents an early legal win for Anthropic, the broader litigation and administrative processes remain unresolved and the case is ongoing.
Contextual note - The court action touches directly on the intersection of national security policy, government procurement, and the public positions of technology providers regarding permissible uses of advanced AI systems.