Two US courts recently reviewed the blacklisting of AI company Anthropic. A judge in the Northern District of California ruled the action illegal, finding Anthropic does not meet the supply-chain risk definition limited to adversary sabotage. However, the DC Circuit Court of Appeals ruled differently, finding the blacklisting permissible under a broader statute, 41 U.S.C. § 4713.

This statute defines supply chain risk more broadly, allowing designation based on risks like data extraction or function manipulation, without requiring proof of malicious intent by an adversary. The DC Circuit affirmed exclusive jurisdiction over the review, effectively reversing the district court's previous ruling. The US designated Anthropic as a supply chain risk under both statutes, with the broader definition encompassing risks like data extraction or function manipulation to surveil or disrupt function.