A US federal judge has issued a ruling declaring that sanctions imposed by the Trump administration against AI firm Anthropic were unlawful. In a 59-page decision released on Thursday, Judge Rita Lin determined that the government had targeted the company as a retaliatory measure for its public criticism of the Pentagon. The ruling effectively makes permanent a temporary suspension of sanctions that had been ordered by the court in March, with the decision taking effect immediately. The government retains the right to appeal the ruling.
The legal action stems from an unprecedented move by defense secretary Pete Hegseth, who had designated Anthropic as a “supply chain risk”—a status typically reserved for foreign entities. This designation, which occurred in February, blocked the company from securing certain military contracts. The move followed Anthropic’s refusal to allow the US military to utilize its AI chatbot, Claude, for autonomous weapons or domestic surveillance. Anthropic executives have contended that this blacklisting could cost the company billions of dollars in lost business and significant reputational harm.
In her decision, Judge Lin, an appointee of Joe Biden, explicitly stated that the government’s invocation of national security cannot serve as a “blank check” to punish those who criticize it. She officially overturned Hegseth’s designation, describing the Pentagon’s action as “illegal and baseless.” The court order also bars the federal agencies named in the lawsuit from enforcing President Donald Trump’s directive to cease the use of Anthropic’s tools.
Anthropic, which maintains that its AI models are not sufficiently reliable for autonomous weaponry and that it opposes domestic surveillance on rights-based grounds, welcomed the ruling. The company filed its lawsuit on 9 March, alleging that the government violated its First Amendment right to free speech by retaliating against its safety-related positions. Additionally, the company argued it was denied due process under the Fifth Amendment, as it was not given an opportunity to dispute the classification. The lawsuit further characterized the government's decision as unsupported by facts and inconsistent with the military’s own previous praise for the Claude models.
The Department of Justice had countered that Anthropic’s refusal to accept specific contractual terms created uncertainty for the Pentagon regarding the operation of military systems, potentially risking the disabling of systems during operations. The government maintained that the designation was a response to contractual disagreements rather than the company’s stance on AI safety. Anthropic is currently facing a separate, pending lawsuit in Washington DC regarding another Pentagon supply-chain risk designation that could potentially exclude the firm from civilian government contracts.
