Pentagon blocked from labeling Anthropic as a ‘supply chain risk'

SAN FRANCISCO (CN) — A federal judge on Thursday found the government likely violated the law when it designated artificial intelligence developer Anthropic a supply chain risk to national security and prohibited military contractors from doing business with it.

“These broad measures do not appear to be directed at the government’s stated national security interests. If the concern is the integrity of the operational chain of command, the Department of War could just stop using Claude. Instead, these measures appear designed to punish Anthropic,” U.S. District Judge Rita F. Lin wrote.

Lin found that there is evidence that Anthropic is being punished for criticizing the government’s position in the press.

“Punishing Anthropic for bringing public scrutiny to the government’s contracting position is classic illegal First Amendment retaliation,” the Joe Biden appointee wrote.

“The record shows that defendants’ conduct appears to be driven not by a desire to maintain operational control when using AI in the military but by a desire to make an example of Anthropic for its public stance on the weighty issues at stake in the contracting dispute,” she added.

The judge also found the Pentagon’s designation of the company as a supply chain risk is likely unlawful because the government had no reason to infer that Anthropic’s questions and concerns about the use of their technology meant they may eventually sabotage it.

“Nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the U.S. for expressing disagreement with the government,” Lin wrote.

Lin’s order prohibits the 17 defendant agencies, including the Department of Defense, the State Department and the Department of Homeland Security from implementing, applying, or enforcing a Feb. 27 presidential directive ordering federal agencies to stop using Anthropic.

It also blocks the Pentagon from implementing, applying, or enforcing Secretary of Defense Pete Hegseth’s Feb. 27 directive labeling Anthropic as a supply chain risk and prohibiting military contractors from doing business with the company, as well as a March 3 letter formalizing the designation.

Lin specified the order “restores the status quo” and does not prevent the defendants from taking any lawful action that they could have taken on Feb. 27, prior to the directives being issued.

“For example, this order does not require the Department of War to use Anthropic’s products or services and does not prevent the Department of War from transitioning to other artificial intelligence providers, so long as those actions are consistent with applicable regulations, statutes, and constitutional provisions,” she wrote.

The order is stayed for seven days to allow the government to appeal.

In a statement, a spokesperson for Anthropic said the company is pleased with the court’s ruling.

“While this case was necessary to protect Anthropic, our customers, and our partners, our focus remains on working productively with the government to ensure all Americans benefit from safe, reliable AI.”

A representative for the defendants did not immediately respond to a request for comment.

Anthropic previously developed AI models to help the Pentagon protect national security and is reportedly its most widely used model. The tech business maintains it always said “Claude,” its flagship family of AI models, should never be used for lethal autonomous warfare and the mass surveillance of Americans. It has no confidence that Claude would be reliable or safe if it supported such warfare, the company says.

However, Anthropic says Hegseth demanded it set aside those restrictions in favor of a new policy. It agreed, except for its two critical principles.

On Feb. 27, after Anthropic said it couldn’t meet the Pentagon’s demand, Trump used a social media post to direct government agencies to stop using Anthropic’s technology, calling it a “radical left, woke company.”

Within hours, the Pentagon designated Anthropic a supply chain risk to national security, with Hegseth posting on X: “Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic … This decision is final."

Anthropic sued the government shortly after, arguing the president’s directive had irreparably harmed the business and the defendants, which include the Pentagon, Treasury Department and several others, want to punish Anthropic for holding onto its views.

At the hearing for a preliminary injunction, the government argued that Anthropic could no longer be trusted to be a partner of the Defense Department because it was raising concerns about how the department conducts military actions and wanted to dictate how its technology was used.

“What happens if Anthropic installs a kill switch and allows its technology to affect war fighters who rely on it in performing missions and operations?” Eric Hamilton, an attorney for the federal government, said.

Lin rejected that argument in her ruling, writing that it was “difficult to understand” how Anthropic could sabotage the defendants by publicly announcing or insisting on restrictions on the use of their technology.

“Defendants appear to be taking the position that any vendor who ‘push[es] back’ on or ‘question[s]’ DoW becomes its ‘adversary.’ That position is deeply troubling and inconsistent with the statutory text,” she wrote.

The judge added that the government violated Anthropic’s due process rights by giving the company no opportunity to respond before labeling it a supply chain risk and accused the Defense Department of “flout[ing] procedural safeguards required by Congress.”

“Anthropic has shown that the supply chain designation was likely contrary to law, both because Anthropic’s conduct did not meet the statutory criteria for a supply chain risk and because the action violated the institutional checks and balances required under Section 3252 and its implementing regulations,” Lin said.

Follow @mhattridge_