Judge: Trump Admin Lacks Evidence for Anthropic Risk Label

A federal judge blocks the Trump administration's attempt to ban artificial intelligence startup Anthropic from securing government contracts, ruling during a Thursday hearing that officials failed to prove the company poses a national security threat. U.S. District Judge Rita Lin rejects the government's designation of Anthropic as a supply-chain risk, marking a major setback for the Department of Defense's efforts to blacklist the technology firm. The decision temporarily halts a federal ban on Anthropic's software, keeping the company's advanced AI models accessible to government agencies.

The legal battle intensifies over stalled contract negotiations between Anthropic and the Pentagon. Anthropic refuses to allow its AI models to be used for mass surveillance of American citizens or for autonomous targeting and firing decisions involving lethal weapons, citing safety concerns and technological limitations. In response, defense officials argue that private corporations should not dictate military operational parameters and assert that all technology integration would comply with lawful standards. Furthermore, government attorneys argue that Anthropic's public criticism of the military justifies the blacklist, a line of reasoning the court finds highly problematic.

This dispute originates from two lawsuits Anthropic filed against the Department of Defense in March, challenging the administration’s restrictive risk designation and subsequent ban. While this specific case proceeds in California, a parallel lawsuit is currently unfolding in a Washington court. The conflict highlights growing friction between Silicon Valley AI developers and the military establishment over the ethical boundaries of artificial intelligence in warfare. In March, the court issued a temporary restraining order against the ban, setting the stage for the current battle over whether to make the injunction permanent.

Defense officials also allege that Anthropic could secretly alter its AI models or trigger a remote "kill switch" to disable military systems during active combat operations. However, technology experts and the court find no evidence to support these claims, noting that once a model is delivered, a developer cannot easily manipulate it. The judge characterizes the government's retaliatory logic as deeply troubling, warning that punishing contractors for voicing ethical concerns sets a dangerous precedent for public-private partnerships.

This ruling carries profound implications for the burgeoning defense tech sector, establishing that the federal government cannot arbitrarily blacklist technology vendors without substantial, verifiable proof of risk. By defending Anthropic's right to set ethical boundaries on its products, the decision protects contractors from administrative retaliation and preserves their freedom of speech. It also signals to other technology firms that they can push back against military demands without fear of immediate, unfounded government bans.

Looking ahead, the court is preparing to issue a final ruling on whether to permanently overturn the administration's ban. As the parallel litigation in Washington continues, the outcome of these cases will likely define the future framework for how the United States military procures and deploys cutting-edge artificial intelligence. Tech companies and defense agencies alike are closely watching the proceedings, which will establish critical legal boundaries for ethical AI deployment in national security for years to come.

Originally reported by TechCrunch