Abstract:
A U.S. federal judge on Thursday blocked the Pentagon from blacklisting artificial intelligence company Anthropic. It's the latest development in a high-stakes dispute between the Claude model developer and the U.S. military over the safety of artificial intelligence on the battlefield. Anthropic filed a lawsuit in California federal court alleging that U.S. Defense Secretary Pete Hegseth exceeded his authority by designating the company as a "national security supply chain risk." Under relevant regulations, the government may make this determination for companies that may expose military systems to the risk of infiltration or destruction by hostile parties.
This rare move by Hegseth previously prevented Anthropic from obtaining some military contracts. The decision comes after Anthropic refused to allow the U.S. military to use the Claude artificial intelligence model for surveillance or autonomous weapons within the United States. Anthropic executives said the move could cost the company billions of dollars in lost business and damage its reputation.
U.S. District Judge Rita Lin, appointed by former Democratic President Joe Biden, said in a 59-page ruling that the Pentagon's decision was "illegal and baseless."
She wrote: "Bland invocations of national security cannot serve as an unlimited authorization to punish and retaliate against government critics."
Anthropic welcomed the ruling and said in a statement that the company remains "focused on working productively with the government to use artificial intelligence to serve national security so that all Americans can benefit from this technology."
The Pentagon did not immediately respond to a request for comment.
Anthropic believes that artificial intelligence models are not yet safe enough to be used in autonomous weapons; at the same time, the company opposes domestic surveillance, saying it violates rights. The Pentagon said private companies should not have the power to restrict military operations.
The Pentagon’s designation is the first time the U.S. government has publicly listed an American company as a supply chain risk under a little-known government procurement regulation designed to protect military systems from foreign sabotage.
Anthropic claimed in its complaint filed on March 9 that the government’s retaliation for its stance on artificial intelligence safety violated the right to free speech guaranteed by the First Amendment to the U.S. Constitution. The company also said the government violated its due process rights guaranteed by the Fifth Amendment to the Constitution before making its determination.
The lawsuit stated that the decision was illegal, lacked factual support, and contradicted the military's positive attitude toward Claude in the past.
The U.S. Department of Justice countered in court documents that Anthropic's refusal to lift relevant restrictions may prevent the Pentagon from determining the scope of Claude's use and may cause military systems to fail during operations.
The government said the finding stemmed from Anthropic's refusal to accept the terms of the contract, not its views on artificial intelligence safety.
Anthropic is also currently filing a second lawsuit in Washington, D.C., over another supply chain risk determination made by the Pentagon. The designation could result in the company being excluded from civilian government contracts.
Comments