17 former security officials urge court to reconsider Pentagon’s Anthropic risk designation

Seventeen former senior national security officials are asking the full D.C. Circuit to reconsider a ruling upholding the Pentagon’s designation of Anthropic as a supply chain risk. Their October 20 brief argues that the decision could discourage AI safeguards and deny federal agencies advanced technology, according to MeriTalk’s October 23 report.

Signatories include former Director of National Intelligence Avril Haines, former national security advisers Jake Sullivan and Susan Rice, and former Deputy Secretary of State Wendy Sherman. The group served in Republican and Democratic administrations, with experience spanning defense procurement, intelligence, cybersecurity and national security policy.

A fight over permitted military uses

The dispute began when the Pentagon sought to use Claude for “all lawful uses.” Anthropic objected to two potential applications: mass surveillance and fully autonomous weapons. After negotiations broke down, Defense Secretary Pete Hegseth directed the Pentagon to designate the company as a supply chain risk. Anthropic challenged that action as unlawful retaliation; the Trump administration defended it as a national security measure.

A three-judge D.C. Circuit panel rejected Anthropic’s challenge in a 2-1 decision under the Federal Acquisition Supply Chain Security Act of 2018. That law gives the government authority to address supply chain risks in federal purchasing. The court found “ample support” for concluding that continued use of Claude by the Defense Department or its contractors presented a national security risk.

The brief challenges the order of decisions

The former officials argue that the Pentagon announced its decision before completing the administrative review required by the supply-chain law. Hegseth issued the formal determination on March 3, four days after publicly announcing the designation. The brief treats that sequence as evidence that the department decided on an outcome and supplied its reasoning afterward.

In the officials’ account, a legitimate determination requires agencies to complete several steps before acting:

• Identify a specific national security threat and assess its potential consequences.
• Consider alternatives to the proposed action.
• Document the reasoning behind the decision.

The filing also alleges retaliation against Anthropic’s public positions on AI safety, in violation of the First Amendment. It cites Hegseth’s criticism of the company’s “sanctimonious rhetoric,” “virtue-signaling” and “Silicon Valley ideology” as evidence that its expressed views influenced the designation. Those are the officials’ allegations, not findings adopted by the appeals panel.

The brief points to a separate August 27 ruling in the Northern District of California granting Anthropic summary judgment on its First Amendment, due process and Administrative Procedure Act challenges. The officials are now asking the full D.C. Circuit to rehear the case and reconsider its panel’s decision.

If the panel decision stands, this Court will validate a sham process that not only weakens national security but gives this and future administrations authorization to punish any company that falls out of favor

The 17 former national security officials, in their October 20 brief, as quoted by MeriTalk

The procurement cost the officials foresee

Beyond the legal objections, the officials warn that excluding Anthropic could weaken federal access to advanced AI as the United States competes with China for technological leadership. Their brief cites the Pentagon’s continued interest in Claude Mythos, which it describes as having advanced cybersecurity capabilities, as an example of technology the government could lose access to through procurement restrictions.

They also argue that the consequences extend beyond one supplier. If safeguards the Defense Department dislikes can invite a supply-chain-risk designation, other developers could be discouraged from establishing restrictions on deployment. The brief frames that prospect as a problem for how AI companies negotiate with the government, not simply a loss for Anthropic.