US Court of Appeals upholds Pentagon's decision to blacklist Anthropic
The 2-1 ruling, issued Friday by the US Court of Appeals for the District of Columbia Circuit, allows the Pentagon to maintain its designation of Anthropic as a national security supply chain risk. The company had challenged the designation, arguing that the government unlawfully targeted it for refusing to remove safeguards against autonomous weapons and mass surveillance.
The decision marks a setback for Anthropic, which says the designation has cost it billions of dollars in lost business and damaged its reputation ahead of a potential initial public offering.
Court Backs Pentagon's National Security Concerns
In the majority opinion, Judge Gregory Katsas, joined by Judge Neomi Rao, said the Defense Department had sufficient grounds to conclude that Anthropic's restrictions could create risks for military operations.
The Pentagon had argued that its military systems must remain available for lawful national security missions without the possibility that a technology provider's built-in safeguards could unexpectedly prevent essential functions.
The court accepted the government's position that Anthropic's refusal to permit unrestricted lawful military use of Claude could undermine operational reliability. It also rejected Anthropic's argument that the designation was retaliation for the company's publicly expressed views on AI safety and ethics.
Judge Karen LeCraft Henderson dissented, disagreeing with the majority's conclusion that the Pentagon had established sufficient legal grounds for the designation.
Dispute Began Over AI Safety Restrictions
The confrontation escalated earlier this year after the Pentagon demanded that Anthropic make Claude available for all lawful military purposes.
Anthropic CEO Dario Amodei refused to remove restrictions preventing the technology from being used in fully autonomous weapons and mass domestic surveillance. The company argued that advanced AI systems are not sufficiently reliable for autonomous lethal operations and that mass surveillance raises serious civil liberties concerns.
Negotiations collapsed, prompting Defense Secretary Pete Hegseth to designate Anthropic a supply chain risk in March. The designation restricted the company's ability to participate in Defense Department contracts and affected contractors using its technology for military work.
The Pentagon maintained that its restrictions were necessary to protect military operations, while Anthropic argued that its safeguards were intended to prevent potentially dangerous applications of its technology.
Separate Court Ruling Complicates the Ban
Friday's decision is separate from a case in California, where a federal judge ruled last month that a parallel government designation against Anthropic was unlawful retaliation for the company's AI safety positions.
The California ruling blocked the broader government-wide ban and restrictions barring military contractors from doing business with Anthropic. The conflicting decisions mean the legal battle over the company's relationship with the government remains unresolved.
Anthropic said it respectfully disagreed with the appeals court's decision and was considering its options, including asking the full appeals court to review the ruling.
The Pentagon and White House did not immediately respond to requests for comment.
