Feds say Hegseth tweet about Anthropic was not a final agency action

SAN FRANCISCO (CN) — Attorneys for the U.S. government told a federal judge Tuesday that a tweet from Secretary of Defense Pete Hegseth designating artificial intelligence developer Anthropic a supply chain risk to national security and prohibiting military contractors from doing business with it “effective immediately,” was not a final agency action.

“Standing here, I’m not aware of any authorities the Department of War has to bar contractors from using Anthropic for non-DoW work,” Eric Hamilton, an attorney for the federal government, said, referring to the Defense Department. “This post itself does not impose obligations on contractors. That flows through supply-chain risk designations.”

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, in a social media post, directed 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, formerly Twitter: “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,” Hegseth concluded the tweet.

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

On Tuesday, U.S. District Judge Rita F. Lin echoed many of Anthropic’s concerns in the opening of a hearing on the plaintiff’s request for a preliminary injunction, saying it seemed the government could have just stopped using Anthropic products, but instead, it decided to go further to punish the company for putting public scrutiny on the department.

“One of the amicus briefs used the term ‘attempted corporate murder.’ I don’t know if it’s murder, but it looks like an attempt to cripple Anthropic,” the Joe Biden appointee said. “There is a concern that Anthropic is being punished for criticizing the government.”

Lin was skeptical of the government’s claim that Hegseth’s tweet should not have been read as effective immediately, given the language of the statement.

“Obviously, the statement is front and center and had a pretty big impact on the public, Anthropic customers and the ecosystem of people who reacted to what happened. You are saying that the sentence ‘effective immediately’ … is just not true? Is that a false statement?” Lin asked.

“I’m saying to read it in context with the sentence preceding it … indicating the DoW will be taking action to do that, which is what it did,” Hamilton responded.

The federal government’s main argument for designating Anthropic a supply-chain risk, Hamilton said, was that the company 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.

He added there were concerns the company may escalate its behavior in the future and manipulate the technology to change how it performs.

“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?” he said.

“That is not a partner the DoW can work with. I can’t think of another scenario where total trust and reliability matter more than this one," he continued.

In rebuttal, Anthropic argued the Defense Department has opportunities to view and test any updates to Anthropic’s technology, and that the company cannot alter any update once it’s deployed.

Hamilton additionally argued that “less intrusive measures,” such as simply stopping the contract with Anthropic and telling contractors not to use the company’s products, would not be sufficient, and that the supply-chain risk designation allows the department to address risks to national security effectively.

“Congress recognized when supply-chain risks exist, the Department of War and other covered agencies don’t have to go contract by contract … it wanted to give the executive the tool with one designation to address a risk,” he said.

However, Anthropic argued the government’s objective was to “generalize the invocation of the supply-chain risk designation,” instead of going through the usual, lawful process of winding down a contract or pursuing suspension if negotiations don’t work out.

“This is something that has never been done to an American company,” Michael Mongan of Wilmer Cutler Pickering Hale and Dorr, an attorney for the Anthropic, said.

“This is not a normal way to respond to the concerns articulated by the other side," Morgan said.

Anthropic argued the Hegseth directive caused it immediate, irreparable harm to its reputation and business interests in the national security sector and beyond.

Mongan said that Anthropic was not asking the court to reinstate the company’s contracts with the government or force the Defense Department to work with Anthropic. Instead, it was requesting a return to the “status quo” on Feb. 27, when the department decided to cut ties with Anthropic.

One way the government could remedy its concerns legally, Mongan said, would be to simply end the contract with Anthropic and transition to another AI model.

“What they can’t do is engage in unconstitutional retaliation for protected speech,” he said. “They can’t impose a secondary boycott that they have conceded has no valid legal basis. They cannot impose an immediate prospective debarment of anthropic for all future govt contracting not supported by any lawful executive authority.”

Lin said she anticipates issuing an order in the next few days.

Representatives for Anthropic did not immediately respond to a request for comment.

A Pentagon spokesperson told Courthouse News that it doesn’t comment on ongoing litigation.

Follow @mhattridge_