A US federal judge has ruled that the Trump administration acted unlawfully in targeting artificial intelligence company Anthropic after the firm refused to allow certain military uses of its AI technology.

US District Judge Rita Lin said the Pentagon’s decision to designate Anthropic as a national security supply-chain risk amounted to unlawful retaliation and violated the company’s constitutional rights.

The dispute began after Anthropic rejected military use of its Claude AI models for mass surveillance and fully autonomous weapons, citing concerns over safety and civil rights. The Pentagon subsequently moved to restrict the company’s access to government and military contracts.

In her ruling, Lin said national security considerations did not give the government unlimited authority to punish a company for expressing concerns about government policy. She also found that Anthropic had been denied procedural protections required under the US Constitution.

The decision represents a major legal victory for Anthropic and could have broader implications for the increasingly important relationship between technology companies and the US military as governments seek access to advanced AI systems.

Anthropic welcomed the ruling and said it remained interested in working with the US government on national security applications of AI. The administration could still challenge the decision through the courts.

The case has also highlighted a growing debate over how far AI companies should be able to restrict military applications of increasingly powerful models, particularly in areas involving surveillance and autonomous weapons.