US judge blocks Pentagon’s Anthropic blacklisting

On Thursday, a U.S. judge halted the Pentagon's decision to blacklist Anthropic, the developer behind the popular AI assistant Claude. Anthropic's legal team argued that Pentagon Secretary Pete Hegseth exceeded his authority when he labeled the company a national security risk, a move that could prevent the company from participating in military contracts. This designation came after Anthropic refused to allow the military to use its Claude AI for surveillance or in the creation of autonomous weapons.

U.S. District Judge Rita Lin, a Biden appointee, ruled that the Pentagon's decision was "illegal and baseless." Judge Lin emphasized that the government's invocation of national security does not grant an unrestricted power to penalize and retaliate against critics. Anthropic expressed satisfaction with the ruling, stating that it remains committed to collaborating with the government to ensure AI benefits all Americans.

The company maintains concerns about the safety and reliability of AI models in autonomous weapons and domestic surveillance.

The Pentagon had not yet provided a response regarding the ruling. This case marks the first instance of a U.S. company being publicly labeled a supply-chain risk under an obscure government procurement statute designed to safeguard military systems from foreign sabotage. Anthropic had filed a lawsuit in March, claiming that the government's actions violated their First Amendment right to free speech and their Fifth Amendment right to due process.

The lawsuit contended that the Pentagon was retaliating against Anthropic's stance on AI safety and that the decision had no factual basis or support. The Justice Department argued that Anthropic's refusal to lift the restrictions posed potential risks to the Pentagon's ability to operate their systems during military operations.