AI Policy Wiki
Dashboard
⚲

Pete Hegseth

high confidence · updated 2026-09-27

US Secretary of Defense (since Jan 2025); central figure in the Anthropic-Pentagon dispute. January 12, 2026 memo demanded renegotiation of existing AI contracts on 'any lawful use' terms and argued 'the risks of not moving fast enough outweigh the risks of imperfect alignment.'

Pete Hegseth is the US Secretary of Defense, confirmed in January 2025. A former Fox News host and Army National Guard major, his most documented AI-policy involvement is the dispute between the Department of Defense and Anthropic over restrictions in Anthropic's government contracts, which began with a memo he issued on January 12, 2026.

Background and role

Before his appointment, Hegseth was a Fox News host and served as a major in the Army National Guard. He was confirmed as US Secretary of Defense in January 2025.

Anthropic contract dispute

The dispute originated with a memo Hegseth issued on January 12, 2026, which demanded renegotiation of existing AI contracts on "any lawful use" terms — removing Anthropic's prohibitions on autonomous weapons and mass surveillance of Americans — and argued that "the risks of not moving fast enough outweigh the risks of imperfect alignment." This memo is the originating primary text of the dispute (Source: theverge.com). The memo pushed to renegotiate Anthropic's Claude Gov contracts to permit "all lawful use," language that Verge reporting confirms originated in the January memo. A Pentagon chief technology officer, quoted in Breaking Defense in February 2026, characterized Anthropic's restrictions as undemocratic.

The Department of Defense set February 27, 2026 as a compliance deadline for Anthropic. On the same date, it designated Anthropic a supply chain risk, a classification typically reserved for foreign adversaries or compromised vendors.

The August 27, 2026 summary-judgment opinion establishes the sequence from the administrative record. At a February 24, 2026 meeting with Dario Amodei and Anthropic's head of policy, Hegseth said Claude had "exquisite capabilities" and that the company's concerns were "understandable," but that the Department "had many other vendors to choose from that have never raised" them and who "would never hold any veto power over" it, and said nothing about the models being unsafe or subject to compromise. He gave the company until 5:00 p.m. on February 27 to accept "all lawful uses," failing which it would be designated a supply chain risk — or, alternatively, said the Department might invoke the Defense Production Act to designate Anthropic as essential to national security and compel its services without the restrictions. On February 27 he posted on X that Anthropic "delivered a master class in arrogance and betrayal," referred to "the sanctimonious rhetoric of 'effective altruism'" and "corporate virtue-signaling," and ordered that "[e]ffective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic" (Anthropic PBC v. U.S. Department of War — Order on Cross Motions for Summary Judgment (N.D. Cal., August 27, 2026)).

The signed memorandum resolves the dating: Artificial Intelligence Strategy for the Department of War is dated January 9, 2026 on its face and was posted to the Department's media server on January 12. Both passages The Verge attributed to a January 12 memo — the "any lawful use" contract directive and "the risks of not moving fast enough outweigh the risks of imperfect alignment" — appear in the January 9 text, which indicates one instrument rather than two. Its subsection "Clarifying 'Responsible AI' at the DoW — Out with Utopian Idealism, In with Hard-Nosed Realism" states that the Department "must not employ AI models which incorporate ideological 'tuning'" and "must also utilize models free from usage policy constraints that may limit lawful military applications," and gives the Under Secretary of War for Acquisition and Sustainment 180 days to incorporate the "any lawful use" language into every contract procuring AI services (Artificial Intelligence Strategy for the Department of War (Hegseth memorandum, 2026); Anthropic PBC v. U.S. Department of War — Order on Cross Motions for Summary Judgment (N.D. Cal., August 27, 2026)).

On March 26, 2026, Judge Rita Lin issued a preliminary injunction blocking the designation, finding that retaliatory government action against Anthropic was likely unlawful. On August 27, 2026 the court entered summary judgment for Anthropic on the First Amendment, due-process and Administrative Procedure Act counts as against Hegseth and the Department, vacated the supply-chain-risk designation and the contractor-boycott order, and permanently enjoined the challenged actions. The court found that Hegseth "failed to make the required reasoned determination regarding less intrusive measures" under 10 U.S.C. § 3252(b)(2)(B), that his six identical notices to congressional committees contained no discussion of such measures as § 3252(b)(3)(B) requires, and that the risk assessment came from the wrong official under 48 C.F.R. § 239.7304(a). It read his public statements as direct evidence that Anthropic's speech was a substantial motivating factor (Anthropic PBC v. U.S. Department of War — Order on Cross Motions for Summary Judgment (N.D. Cal., August 27, 2026); Anthropic PBC v. U.S. Department of War — Order of Final Relief (N.D. Cal., August 27, 2026)). The case is tracked in Anthropic v. United States (Pentagon ban challenge). Hegseth is the first US Secretary of Defense to take direct public action against a frontier AI lab over its use policies.

The Department's parallel exclusion of Claude under the Federal Acquisition Supply Chain Security Act (41 U.S.C. § 4713), made by Hegseth's written determination of March 3, 2026 and reaffirmed when he denied reconsideration on June 3, was upheld by a divided D.C. Circuit panel on September 25, 2026. The majority (Katsas, J., joined by Rao, J.) held that the Secretary "did not transgress any limits on his authority under the Supply Chain Security Act or the Constitution" and read his February 27 social-media post, which criticized Anthropic's "sanctimonious rhetoric" and "virtue-signaling," as addressed to the company's refusal of an "all lawful uses" contract term rather than to its advocacy; Judge Henderson dissented (Anthropic PBC v. U.S. Department of War — D.C. Circuit Opinion, Nos. 26-1049 and 26-1162 (September 25, 2026); Source: cnbc.com). The opinion records that Hegseth's March 3 determination found that use of Claude "presents a significant supply chain risk," that no "less intrusive measures" were "reasonably available," and that an "urgent national security interest" required immediate action; that in his June 3 order denying reconsideration he clarified the determination "did not depend" on any understanding of Anthropic's "real-time technical access to or control" of Claude after deployment; and that his February 27 post described Anthropic's conduct as a "textbook case of how not to do business" with the Pentagon and its position as a "veto power over the operational decisions of the United States military" (Anthropic PBC v. U.S. Department of War — D.C. Circuit Opinion, Nos. 26-1049 and 26-1162 (September 25, 2026)).

Civilian-harm mitigation and AI targeting

Officials involved in the Pentagon's internal investigation of the February 28, 2026 strike on a school in Minab, Iran told Bloomberg that Hegseth had dismantled most of the department's civilian harm mitigation units, cutting headcount across a number of teams by roughly 90% to fewer than 20 staff, with Central Command's team reduced from 10 to one, and that no member of such a team reviewed the Minab site before the strike; they cited the cuts, a decision not to involve the group in planning, and overreliance on the AI in Maven Smart System among the contributing failures. Some of Hegseth's advisers have expressed concern that political scrutiny of Maven could disrupt its further adoption, according to the same officials (Source: bloomberg.com). A March 12, 2026 letter from House Democrats led by Reps. Sara Jacobs, Yassamin Ansari and Jason Crow had asked Hegseth whether AI, including Maven, was used to identify the school as a target, and recorded his March 2 statement that there would be "No stupid rules of engagement" House Democrats' letter to Secretary Hegseth on civilian harm in Operation Epic Fury (2026). CNN's September 18, 2026 account of a false AI-assisted intelligence report that nearly led to the boarding of a Chinese ship cited the January 9, 2026 AI strategy memorandum's direction to put AI models "directly in the hands of our three million civilian and military personnel, at all classification levels" (Source: cnn.com; Artificial Intelligence Strategy for the Department of War (Hegseth memorandum, 2026)). Sens. Mark Warner, Christopher Coons and Jack Reed wrote to Hegseth and Director of National Intelligence Jay Clayton in a letter dated September 19, 2026 asking them to give the relevant Inspectors General unrestricted access to investigate the Minab strike, the aborted SOCPAC interdiction and any unreported errors in AI-enabled targeting workflows. The letter describes the department's AI strategy as treating AI governance policies as "blockers" and refers, without naming the company, to "the Department of Defense's harassment of a leading U.S. AI vendor over the latter's reluctance to allow its products' use in fully-autonomous kinetic operations" (Sens. Warner, Coons and Reed — letter to Secretary Hegseth and DNI Clayton on AI targeting and intelligence errors (2026)). See DOD — Department of Defense (AI Deployer).

Relationships

Sources

  • Anthropic Claude Gov + Pentagon Dispute (2025–2026)
  • Reuters / Exclusive: Pentagon clashes with Anthropic (Jan 29, 2026)
  • Axios / Pentagon threatens to cut off Anthropic (Feb 15, 2026)
  • Breaking Defense / Pentagon CTO says 'not democratic' for Anthropic to limit military use (Feb 2026)