Washington: Anthropic filed a federal lawsuit against the Trump administration on Monday, claiming the government illegally retaliated against the artificial intelligence firm after it barred its "Claude" model from being utilised for autonomous lethal combat and domestic mass surveillance.

The 48-page complaint, lodged in U.S. District Court in San Francisco, seeks to overturn and block a recent government designation labelling the company a national security supply-chain risk. Anthropic, a leader in the AI sector, argued that the move was a direct punitive response to its refusal to abandon its core safety principles.

The company stated in the filing that it was established on the conviction that AI must be "used in a way that maximises positive outcomes for humanity" and should "be the safest and the most responsible."

"Anthropic brings this suit because the federal government has retaliated against it for expressing that principle," the lawsuit states.

The designation marks the first time a domestic American firm has been publicly sanctioned with such a label, a status typically applied to entities from hostile foreign nations, such as the Chinese telecommunications giant Huawei. Under the order, all defence contractors and vendors must certify they have excluded Anthropic’s models from their operations with the Pentagon.

The legal challenge describes the financial stakes as "enormous," alleging the administration is "seeking to destroy the economic value created by one of the world's fastest-growing private companies."

The conflict reached a breaking point after Anthropic’s leadership drew the ire of Defence Secretary Pete Hegseth. The company insisted its technology remain restricted from use in fully autonomous weapons platforms and large-scale surveillance of American citizens. In response, President Donald Trump issued a directive for all federal agencies to terminate the use of Anthropic’s technology.

Shortly thereafter, Hegseth officially classified the firm as a "Supply-Chain Risk to National Security," prohibiting military suppliers and partners from engaging in "any commercial activity with Anthropic." A six-month transition period was granted for the Pentagon’s internal systems.

The timing of the blacklisting is particularly sensitive, occurring just days before the U.S. military launched strikes against Iran. Claude currently serves as the Pentagon’s most utilised frontier AI model and remains the only such system integrated into the Department of War’s classified networks.

In its legal arguments, Anthropic alleges the administration’s actions violate the First Amendment by penalising the company for its speech regarding AI safety. The suit further contends the Pentagon exceeded its statutory authority and denied the company due process as guaranteed by the Fifth Amendment.

"The Constitution does not allow the government to wield its enormous power to punish a company for its protected speech," the complaint says.

Anthropic is asking the court for a permanent injunction to prevent the enforcement of the national security designation.

With inputs from AFP