A federal judge has ruled that the US Department of Defense illegally retaliated against Anthropic after the artificial intelligence company refused to remove safety restrictions on its Claude large language model for military uses including mass surveillance and lethal autonomous warfare.
In a 59-page order, US District Judge Rita Lin granted Anthropic summary judgment on First Amendment, due process and Administrative Procedure Act claims. The court found the Pentagon’s “supply chain risk” designation was imposed without the pre-deprivation process required by the Fifth Amendment, outside the supply chain statute, and was arbitrary and capricious. Judge Lin wrote that “the empty invocation of national security is not a blank check to punish and retaliate against government critics.”
The dispute escalated after negotiations broke down over Anthropic’s refusal to accept a clause permitting “all lawful uses.” On February 27, former President Donald Trump ordered federal agencies to stop using the company’s technology, calling it a “RADICAL LEFT, WOKE COMPANY” on Truth Social. Defense Secretary Pete Hegseth then barred military contractors from doing business with Anthropic. In March, the Pentagon designated Anthropic a supply-chain risk, a first for a US company.
The government later abandoned its central risk assessment claim and conceded that Anthropic has no backdoor access to deployed models and that Claude is no riskier than any other “black box” AI system. Judge Lin noted the timeline showed the designation followed public criticism rather than legitimate security concerns. She vacated the designation and Hegseth’s boycott order, issued a permanent injunction, and denied the government’s request for a seven-day administrative stay.
Anthropic still faces a parallel case in Washington, D.C., meaning its security-risk status is not fully cleared. The company said it welcomed the ruling and remains focused on working productively with the government on national security applications. The decision could set a precedent for how federal agencies handle AI developers that impose ethical limits on military use.
Separately, Anthropic announced the Model Hardware Standard research preview, aimed at helping AI assistants such as Claude connect with robots and specialized laboratory and factory equipment.
