Judge temporarily blocks Trump administration’s ban: NPR

A federal judge in San Francisco on Thursday issued a preliminary injunction against the Pentagon, temporarily freezing its potentially disastrous decision to classify Anthropic as a “supply chain risk.” The company and the Department of Defense are in the midst of a dispute over how the military uses the company’s artificial intelligence model, Claude.

Judge Rita F. Lin of the Northern District of California also temporarily blocked President Trump’s directive ordering all federal agencies to stop using Anthropic’s technology.

These measures suspend the government’s ban until a court decides the merits of the underlying litigation.

In his order, Lin wrote that supply chain risk designations are typically limited to foreign intelligence agencies and terrorists, not U.S. companies. “These far-reaching measures do not appear to be directed at the government’s stated national security interests,” Lin wrote. “If the concern was the integrity of the operational chain of command, the Department of the Army could simply discontinue use of the Claude.”

“Rather, these measures seem aimed at punishing humanity,” she continued.

The injunction stems from a contract dispute between Anthropic and the Department of Defense that was made public in February and escalated in the weeks since.

Anthropic CEO Dario Amodei said the company would not allow Claude to be used for autonomous weapons or surveillance of American citizens. The Pentagon has said it is up to the military, not companies, to decide how to use tools purchased from contractors.

President Trump ordered all federal agencies to stop using Claude and increased the terms.

pentagon specified Earlier this month, citing national security, it cited humanity as a “supply chain risk.” In a statement at the time, the military said: “We will not allow vendors to restrict the lawful use of critical capabilities and enter the chain of command, putting our warfighters at risk.”

Anthrop filed two lawsuits in federal court alleging the designation amounted to unlawful retaliation. that stance It also addressed the safety of AI, arguing that the label would harm both customers and revenue because it would also prohibit Department of Defense contractors from doing business with the company. The lawsuit also alleges that the Trump administration violated the company’s First Amendment speech rights.

a wide range of organizations including microsoft, ACLU and Veteran Mentorsubmitted amicus Court brief in support of Anthropic.

At Tuesday’s hearing, Lin seemed heavily inclined to grant the preliminary injunction, saying his first impression was that Antropic’s ban looked like punishment for openly opposing the government’s position.

In court, Defense Department lawyers argued that Anthropic’s actions made it unreliable and that the supply chain risk designation stemmed from the company’s decision to thwart the military’s use of its AI models, not from publicly opposing the Department’s position on the issue.

They also argued that, in theory, the company could update Claude in a way that would endanger national security.

In his order, Lin said the supply chain risk designation “is in violation of the law and is likely to be arbitrary and capricious.”

She wrote that there is nothing in the statute to apply supply chain risk designations that supports “the Orwellian idea that U.S. companies could be branded potential adversaries and disruptors of the United States for exposing their disagreements with the government.”

She also wrote that the Pentagon praised Anthropic as a partner and conducted rigorous national security reviews. But until the company publicly raised concerns about how its technology might be used, the Pentagon “announced plans to cripple Anthropic, blacklisting it from doing business with any company that serves the U.S. military, permanently cutting off its ability to cooperate with the federal government, and branding it an adversary with the potential to engage in sabotage,” she writes. [the Department of War] And that introduced supply chain risks. ”

“This seems like classic First Amendment retaliation,” she continued.

Mr Antropic welcomed the judge’s decision. “We are grateful to the court for moving quickly and are pleased that the court agreed that Anthropic is likely to succeed on the merits. While this litigation was necessary to protect Anthropic, our customers, and partners, we remain focused on working productively with the government to ensure all Americans receive the benefits of safe and reliable AI,” a spokesperson wrote in an email to NPR.

The Pentagon did not immediately respond to a request for comment, but previously told NPR it has a policy of not commenting on ongoing litigation.

Jennifer Huddleston, senior technology policy fellow at the Cato Institute, a libertarian think tank, said the preliminary injunction reads as if the judge believes Anthropic has a good chance of success on the merits.

Mr Huddleston said the decision was important and had far-reaching implications beyond this particular case.

“This preliminary injunction really goes into some of the typical issues, like ensuring that businesses and individuals are not subject to retaliation for exercising their First Amendment rights, and ensuring that there is appropriate due process when such important decisions are made in matters that could potentially harm a business,” she said.