A three-judge DC Circuit panel ruled 2-1 on 25 September that the Defense Department can keep its supply-chain risk designation for Anthropic under section 4713. The majority tied the decision to Claude's encoded limits on lethal autonomous warfare and domestic mass surveillance. A California court previously found a parallel designation unlawful under a different provision, section 3252; the new appeal ruling concerns the separate Pentagon action.
The US Court of Appeals for the DC Circuit denied Anthropic's petition on 25 September in a 2-1 decision. The ruling lets the Defense Department maintain its designation of the company as a supply-chain risk under section 4713. Judge Gregory Katsas wrote for the majority, joined by Judge Neomi Rao; Judge Karen Henderson dissented.[1], [2]
The opinion describes limits Anthropic built into its AI model Claude against lethal autonomous warfare and mass domestic surveillance. Those limits had blocked tasks requested by government users more than once. The majority treated the resulting supply-chain question under section 4713, which it said does not require proof that the supplier acted with a bad motive.[1], [2]
A Northern California federal court had found a parallel designation unlawful under section 3252, a provision the DC Circuit majority distinguished because it does require bad motive. The September appeal ruling addresses the separate section 4713 action. Anthropic said it disagrees with the new decision and is considering further review, while pointing to the California ruling.[1], [2]
