US court rules ban of Anthropic unlawful, punished for ‘publicly criticising’ Pentagon

A US federal judge has ruled that sanctions imposed on AI firm Anthropic by the Trump administration were unlawful. The court found that the Pentagon's February order against Anthropic was intended to punish the company for publicly criticizing the government. Judge Rita Lin, in her 59-page decision, emphasized that the use of national security as an excuse to retaliate against government critics is not a justification for punishment.

The judge vacated the company's supply-chain risk designation, stating that the Department of Defense's actions violated the due process clause of the Fifth Amendment. The court ordered the Department of Defense to revoke all guidance, directives, communications, or instructions targeting Anthropic. In February, the Pentagon blacklisted Anthropic, the maker of Claude, due to the company's refusal to allow the military to use the AI chatbot for US surveillance or autonomous weapons.

Anthropic argued that AI models are not safe for use in autonomous weapons and oppose domestic surveillance as a violation of rights. The decision also blocked Anthropic from certain military contracts, potentially costing the company billions in lost business and reputational harm. Anthropic is currently involved in a second lawsuit in Washington, DC, over a separate Pentagon designation that could lead to its exclusion from civilian government contracts.