DC Circuit Upholds Pentagon Blacklisting of Anthropic in 2-1 Ruling

The US Court of Appeals for the DC Circuit issued a 2-1 decision approving the Pentagon's exclusion of Anthropic from military contracts under §4713 of the Supply Chain Security Act. The court found that Claude's built-in safety restrictions qualify as a supply-chain risk because they could cause the model to refuse government-requested tasks and create operational uncertainty during military operations. The core dispute was Anthropic's refusal to accept the Pentagon's 'all lawful use' standard for AI deployment, which other model-makers agreed to. Anthropic drew specific red lines against use of its tools for mass domestic surveillance or autonomous weapons. The court acknowledged competing risks: constrained AI models potentially causing military operations to fail versus unconstrained models hallucinating inappropriate targets for lethal military force. The court accepted Anthropic's AI-safety advocacy as protected speech but held the contract exclusion flowed from rejected contract terms rather than retaliation for that advocacy. A separate California court, Judge Rita Lin, reached a different outcome on a related but distinct designation under §3252, issuing a preliminary injunction on three independent grounds: First Amendment retaliation, Fifth Amendment due process, and APA violations. Judge Lin also found the government relied on the wrong official's risk assessment and had not considered less intrusive measures. The DC Circuit said it had 'no quarrel' with the California court's conclusion, stressing that §4713 carries a substantially broader definition of supply-chain risk than §3252.