A federal judge in San Francisco ruled on Thursday that sanctions the Trump administration imposed on artificial intelligence company Anthropic in February were illegal, finding the government had retaliated against the firm for publicly criticising the Pentagon.
In a 59-page decision, Judge Rita Lin wrote that the blacklisting violated Anthropic’s First Amendment rights. “The empty invocation of national security is not a blank check to punish and retaliate against government critics,” she stated.
The dispute traces to a public clash between Anthropic and the Pentagon over how the US military uses artificial intelligence. The company had raised objections to certain military applications of AI, positioning itself against the administration’s approach.
The government responded in February by placing the AI developer on a blacklist, a measure that restricted its dealings with federal agencies. The court found that punitive step was tied directly to the company’s public statements rather than any legitimate security concern.
Anthropic, maker of the Claude family of AI models, has grown into one of the most prominent US startups in the sector and has repeatedly found itself at odds with the current administration over policy and oversight.
The ruling forms part of a broader pattern of legal setbacks for the administration. Separately on Thursday, a judge placed a 14-day hold on the government’s implementation of an order to limit mail-in voting, with a swift appeal considered likely.
The decision reinstates Anthropic’s standing to work with federal bodies and rejects the argument that the blacklisting served a protective purpose. The court framed the measure as an act of reprisal against a critic.
The Trump administration has not indicated whether it will appeal the ruling on the Anthropic sanctions.