A federal judge in California has temporarily blocked the U.S. Department of Defense’s decision to list Anthropic, the developer of the Claude AI model, as a national security and supply chain risk, in a new twist in a high-profile conflict between the military and a technology company over the safety of artificial intelligence on the battlefield.
A federal judge in California has temporarily blocked the U.S. Department of Defense’s decision to list Anthropic, the developer of the Claude AI model, as a national security and supply chain risk, in a new twist in a high-profile conflict between the military and a technology company over the safety of artificial intelligence on the battlefield.
The conflict arose after US Defense Secretary Pete Hegseth declared Anthropic a defense procurement risk. This status, usually applied to companies that expose military systems to the risk of foreign espionage or sabotage, deprived the developer of some government defense contracts.
The Pentagon's move was prompted by Anthropic's refusal to allow the use of its Claude AI models for military surveillance and autonomous weapons. Anthropic said the significant restrictions were due to security concerns — AI technologies are currently not reliable enough for autonomous use in weapons, and domestic surveillance violates citizens' rights. In response, the Defense Department said that private companies should not dictate restrictions on military operations.
In March, Anthropic filed a lawsuit, saying the Pentagon's decision was baseless and could cost the company billions of dollars in reputational and financial damages. The developer's lawyers cited violations of the First Amendment of the US Constitution (retaliation for a position on AI safety) and the Fifth Amendment (violation of due process).
Federal Judge Rita Lin ruled the Pentagon's decision illegal and unreasonable in a 59-page ruling.
"Empty references to national security are not carte blanche to punish and crack down on government critics," the judge emphasized in her ruling.
Anthropic welcomed the court's ruling and said it looks forward to continuing to work constructively with the government to use AI for national security. The US Department of Defense declined to comment.
At the same time, this is not the developer's only legal battle: another lawsuit by Anthropic against the Pentagon is being considered in Washington due to similar restrictions that could block the company's access to US civilian government contracts.
