Anthropic sues Pentagon over killer AI rift

The tech firm was blacklisted for resisting the US War Department’s demand to lift limits on surveillance and autonomous weapons

AI developer Anthropic has filed a lawsuit against the Pentagon and Secretary of War Pete Hegseth, challenging the Trump administration’s decision to designate the company a national security risk and blacklist its technology across the federal government.

The complaint, filed Monday in the US District Court for the Northern District of California, argues that the administration exceeded its legal authority and retaliated against the company for refusing to remove safeguards governing how its AI systems can be used by the military.

Anthropic is asking the court to declare the government’s actions unlawful and block the measures, which it says threaten to cause “irreparable” harm to one of the world’s “fastest-growing” and most “responsible” AI companies.

Read more

The legal battle stems from the company’s alleged refusal to lift restrictions preventing its models from being used for mass domestic surveillance or fully autonomous weapons.

On February 27, US President Donald Trump ordered federal agencies to stop using Anthropic’s technology, accusing it of being run by “leftwing nut jobs.” Soon afterward, Hegseth formally designated the company a “supply-chain risk to national security,” a label normally reserved for firms linked to foreign adversaries.

The administration has argued that the company’s refusal to modify its models so they can be used for “all lawful military purposes” poses operational risks by giving a private contractor leverage over battlefield decision-making. Anthropic’s lawsuit disputes that characterization, saying the administration’s actions amount to punitive retaliation over a policy disagreement about AI governance.

Read more

The company was previously the only AI developer authorized to deploy large language models on classified Pentagon networks, where its Claude system supported intelligence analysis and operational planning under contracts worth up to $200 million.

Claude also reportedly played a key role in planning and conducting the raid targeting Venezuelan President Nicolas Maduro – as well as the US-Israeli strikes on Iran launched mere hours after Trump blacklisted the company.

Under the administration’s order, federal agencies were instructed to phase out the technology within six months while transitioning to alternative AI providers. Anthropic said the continued reliance on its systems during that period undermines the government’s claim that the company represents an immediate security threat.