U.S. Appeals Court Upholds Pentagon Labeling Anthropic a Security Risk - WSJ

A federal appeals court ruled on Friday that the Pentagon may designate the artificial‑intelligence company Anthropic as a national security risk, a determination that could bar the firm from future Department of Defense contracts. The decision upholds a request by the Trump administration to blacklist the firm on security grounds.

The court’s order affirms the Pentagon’s authority to assess and act on perceived threats posed by technology providers, including those that develop advanced generative‑AI systems. By labeling Anthropic a security risk, the department can exclude the company from participating in new procurement opportunities, though the ruling does not automatically terminate any existing agreements.

Anthropic, known for its Claude series of language models, has been under scrutiny for potential vulnerabilities related to data handling and possible foreign influence. The Pentagon’s assessment cited concerns that the firm’s technology could be leveraged in ways that compromise U.S. defense interests.

Legal analysts note that the appellate court’s affirmation reinforces the executive branch’s discretion in safeguarding national security through procurement controls. The decision also signals that similar designations could be applied to other AI developers if the department deems it necessary.

The ruling marks a significant step in the government’s broader effort to regulate emerging AI technologies and ensure that suppliers meet security standards before gaining access to defense contracts.