Appeals court upholds Pentagon’s Anthropic blacklist

A federal appeals court ruled 2-1 that the Pentagon may continue treating artificial intelligence company Anthropic as a supply-chain risk, allowing the department to remove its Claude models from government systems and restrict their use in defense work.

The U.S. Court of Appeals for the District of Columbia Circuit found that the defense secretary acted within his authority under the Federal Acquisition Supply Chain Security Act. The dispute arose after Anthropic declined to permit Claude to be used for fully autonomous weapons or mass surveillance of Americans.

Writing for the majority, Judge Gregory Katsas said the statutory definition of a supply-chain risk “turns on what Anthropic does, not why Anthropic does it.” The court said officials had to balance the risk that constrained AI could fail during military operations against the danger that unconstrained systems could hallucinate targets for lethal force. Judge Neomi Rao joined Katsas in the majority.

The ruling contrasts with a decision by a federal district court in San Francisco, which found that parallel government actions against Anthropic amounted to unlawful First Amendment retaliation. The cases involved different statutes and avenues for judicial review.

Anthropic said it disagreed with the D.C. Circuit’s decision and was considering further review. The company could seek reconsideration by the full appeals court or petition the Supreme Court.