Kathmandu— A US federal judge has struck down the Pentagon’s designation of AI developer Anthropic as a supply-chain risk to national security, finding that the move was unlawful retaliation against the company for voicing concerns about the use of its technology. Judge Rita Lin ruled on Thursday that Secretary Pete Hegseth overstepped his authority and violated Anthropic's First and Fifth Amendment rights by blacklisting the firm after it objected to using its Claude AI model for mass domestic surveillance or autonomous weapons systems. The decision follows a lawsuit filed by Anthropic in March, alleging the Pentagon retaliated against protected speech and denied due process.

Escalating Dispute with the Pentagon

The dispute originated from the Pentagon’s desire to integrate AI models like Claude into military operations for “any lawful use.” Anthropic had previously collaborated extensively with the Pentagon, even securing an agreement potentially worth up to $200 million. However, CEO Dario Amodei expressed reservations in February about deploying the model for applications he deemed unreliable or ethically questionable, specifically mass surveillance and fully autonomous weapons systems.

Government Retaliation and Blacklisting

Following Amodei’s statements, President Donald Trump directed federal agencies to stop using Anthropic technology on February 27, labeling the company a “radical Left AI company.” Secretary Hegseth subsequently announced Anthropic's designation as a supply-chain risk, effectively barring other Pentagon partners from commercial dealings with the firm. Anthropic responded by filing suit, arguing that this treatment was historically reserved for US adversaries and violated its constitutional rights.

Court Finds Overreach of Authority

In her 59-page ruling, Judge Lin determined that Hegseth’s actions were “arbitrary and capricious” and exceeded his authority. She emphasized that the government's invocation of national security could not justify punishing a company for expressing critical views. According to the judge, the Pentagon sought to make an example of Anthropic for what it perceived as arrogance in challenging Washington’s policies. Lin clarified that while the Pentagon is free to choose its AI vendor and abandon Claude if necessary, the manner in which it attempted to punish Anthropic was unlawful.

Pentagon Reluctance to Lose Access

Despite the conflict, defense officials acknowledged the value of Anthropic's technology. One official told Axios in February that “the problem for these guys is they are that good.” The ruling underscores the Pentagon’s freedom to select its preferred AI vendor but limits its ability to retaliate against companies expressing concerns about specific applications.

Anthropic also has a separate lawsuit pending in Washington, DC concerning another supply-chain risk designation that could impact its eligibility for civilian government contracts.

(With inputs from RT)

Originally published on abcnews.com.np.