Microsoft asks court to block Trump Pentagon ban on Anthropic AI

Body

Microsoft and retired military leaders backed Anthropic’s effort to block a Trump administration move that excluded the artificial-intelligence company from military work after the government labeled it a supply-chain risk.

In a Tuesday filing tied to a federal lawsuit Anthropic filed on Monday, Microsoft said the Defense Department’s decision—made by Defense Secretary Pete Hegseth—was driven by a contracting dispute rather than legitimate national-security interests. Microsoft argued that using a supply-chain risk designation to resolve a contractual disagreement could have “graves effects economic that do not serve the public interest,” according to the company’s court submission.

Microsoft said the Defense Department’s action would compel government contractors to follow “vague and mal defined” guidelines that, he said, had not previously been applied publicly to a U.S. company. The company asked a judge to order a temporary lifting of the designation so Anthropic and the Trump administration could have what Microsoft described as a “more reasoned” discussion.

The filing also responded to the friction that has surrounded Anthropic’s refusal to allow unrestricted military use of its Claude model. Microsoft said it agreed with Anthropic’s two “ethical red lines” that became points of contention in negotiations after the Pentagon insisted Anthropic allow “all legal uses” of its AI. Microsoft further told the court that it believes American AI should not be used for “mass surveillance” within the country or to start a war without meaningful human control, and that it said the position has broad support in the United States, as the government recognized.

Anthropic’s case in San Francisco is before U.S. District Judge Rita Lin, who was nominated by President Joe Biden in 2022. Lin scheduled a hearing for March 24. The company also filed a separate, narrower case in federal appeals court in Washington, D.C.

The Monday-to-Tuesday filings were bolstered by additional court submissions, including one from a group of 22 retired senior U.S. officials from across the services, some of whom had been secretaries of the Air Force, the Army, and the Navy, and a retired Coast Guard command leader. That submission argued that Hegseth’s conduct was misuse of government authority as “represalia” against a private company that had “disgusted the leadership,” and it warned that the secretary’s actions threatened principles of the rule of law that have long supported U.S. forces.

The military-leader support included the presence of former CIA director Michael Hayden, a retired Air Force general, and retired Coast Guard Admiral Thad Allen, who led the U.S. government response to Hurricane Katrina. The retired officials told the court that instead of protecting U.S. national security, the secretary’s conduct threatened rule-of-law principles, according to their filing.

Although the dispute between the Trump administration and Anthropic was reported to have begun in close timing with announcements by Trump and Hegseth and unfolded amid a broader security environment, the court papers described uncertainty about how rapidly withdrawing access to an AI technology integrated into military platforms could affect ongoing planning and operational safety. The filings did not mention the Iran war that began shortly after the administration’s actions.

Separately, the current commander of U.S. Central Command confirmed in a video posted Wednesday on social media that the U.S. military used “advanced AI tools” to “examine large amounts of data in seconds,” while also saying the final decisions on targeting remained human. In the video, Vice Adm. Brad Cooper said such tools allow leaders to make decisions with greater speed, while emphasizing that humans still decide which targets to engage and when.

The Pentagon declined to comment, and said it does not comment on matters in litigation. The Administration’s decision leaving Anthropic out of military work was previously described as the result of a dispute over how it would be used, and military officials have said efforts were underway to shift that work to competitors such as Google, OpenAI and xAI.