Court rules Trump can blacklist Anthropic for refusing to enable Claude features
The U.S. Court of Appeals for the D.C. Circuit upheld the government’s authority to blacklist Anthropic under a broad procurement statute, while not disputing a separate Northern District of California finding that Anthropic did not meet a narrower statutory definition of a supply-chain risk. The appeals court said it has exclusive jurisdiction to review designations under 41 U.S.C. § 4713, which does not require a showing of malicious intent.
Why It Matters
The decision preserves a government pathway to restrict a major AI firm even though a different court found Anthropic did not qualify as a threat under a narrower statute; the split highlights how choice of legal authority can shape oversight of AI products used by the military. Judges also raised concerns that overly constrained AI models could jeopardize military operations, signaling operational risks factored into the legal review.
Key Facts
- Courts involved: U.S. District Court for the Northern District of California; U.S. Court of Appeals for the D.C. Circuit
- Earlier district court finding: Northern District ruled last month that the blacklisting was illegal under 10 U.S.C. § 3252 because Anthropic did not meet the statute's definition of supply-chain risk
- Appeals court ruling: D.C. Circuit did not dispute the district court's primary finding but said it has exclusive jurisdiction to review the designation under 41 U.S.C. § 4713
- Statutes at issue: 10 U.S.C. § 3252 (narrower, requires adversary/malicious action) and 41 U.S.C. § 4713 (broader procurement authority)
- Government designation: The U.S. designated Anthropic as a supply-chain risk under both § 3252 and § 4713
The D.C. Circuit on appeal addressed the U.S. government's blacklisting of Anthropic, ruling that review of the designation under 41 U.S.C. § 4713 falls within the court's exclusive jurisdiction. The appeals court did not overturn the Northern District of California's earlier finding that Anthropic did not meet the narrower statutory test for a supply-chain risk under 10 U.S.C. § 3252, which focuses on adversarial or malicious actions.
The D.C. Circuit emphasized the different statutory language and scope between the two laws. The district court had concluded that the word 'adversary' and the surrounding terms in § 3252 — including sabotage and malicious introduction of unwanted functions — imply that bad motive is required to support a designation. The appeals court agreed with that interpretation for § 3252 but noted that § 4713 uses a much broader definition of supply-chain risk and does not contain the same limitation.
Under § 4713, the court noted, 'supply chain risk' covers the risk that any person may sabotage, maliciously introduce unwanted function, extract data, or otherwise manipulate various aspects of covered technology products so as to surveil, deny, disrupt, or otherwise manipulate their function or the information they handle. Because the government had designated Anthropic under both statutes, the appeals court focused on whether the § 4713 designation was subject to its exclusive review.
The judges also raised operational concerns tied to constrained AI systems, observing that overly limited models could imperil military operations. While the appeals court did not contest the district court’s factual finding about Anthropic’s intent, it concluded that the broader procurement statute provides a separate, permissible basis for the government's action and that challenges to such procurement-based designations fall within the D.C. Circuit’s authority.
Keep Reading
Huge fire on vehicle ferry off Mykonos prompts mass evacuation
Houthi attack on Mecca, Medina would cross ‘red line’, Pakistan PM tells UN
Italy crisis deepens as Belgium’s win stalls World Cup rebuild in Rome
