Federal court rules Pentagon's action against Anthropic was unlawful retaliation
A federal judge has ruled that the Pentagon illegally penalized AI company Anthropic for its criticism of the Defense Department's AI policies. The judge found the government's supply chain risk designation was based on a desire to punish Anthropic for 'arrogance' rather than any real security threat. Anthropic had sued after being accused of endangering national security over its refusal to allow unrestricted military use of its technology.
Judge Rita Lin's 59-page ruling determined the Pentagon's supply chain risk designation stemmed from Anthropic's public criticism rather than verifiable security concerns. The dispute originated in February when Defense Secretary Pete Hegseth accused the company of endangering national security after CEO Dario Amodei resisted allowing its AI models to power mass surveillance or autonomous weapons systems.
The ruling follows earlier interim protections Lin granted, including blocking enforcement of a White House directive barring federal agencies from using Anthropic's Claude chatbot. A separate legal challenge remains pending in Washington's federal appeals court, while competitor OpenAI secured its own Pentagon partnership within hours of the government's initial action against Anthropic.
This ruling could reshape how technology companies engage with federal agencies on politically sensitive matters. If upheld, it may encourage other firms to publicly challenge government policies without fear of contractual retaliation, potentially altering the balance of power in government-contractor relationships. The case also highlights growing tensions between national security priorities and ethical constraints on emerging technologies, with implications for how AI developers navigate military applications while protecting civil liberties.