US court sides with Pentagon in Anthropic AI ban

The US appeals court ruled in favor of the Pentagon on Friday, upholding its decision to ban Anthropic from its supply chain. This decision came after Anthropic refused to remove certain safeguards from its Claude AI model, which the Pentagon wanted to prohibit from being used for fully autonomous lethal weapons and mass domestic surveillance of Americans.

The conflict began in February when Anthropic rejected Pentagon Secretary Pete Hegseth's demand to remove restrictions on using Claude for these purposes. As a result, the military canceled Anthropic's contracts and barred other Pentagon contractors from using its technology. An Anthropic executive had even raised concerns about Claude's use by a Pentagon contractor, Palantir, during a military operation that captured Venezuelan President Nicolas Maduro on January 3.

The appeals court opinion, written by Judge Gregory Katsas and representing two of the three judges, stated that "in our Republic, it is the President and the Secretary of War who must determine how best to balance risks." This marks an unusual application of authority against a US company. In response, Anthropic's competitor OpenAI signed an agreement with the Pentagon on the same day, while other rivals like xAI, Google, and Microsoft followed soon after.

Anthropic, valued at nearly one trillion dollars and preparing for a major IPO, has since become a pariah at the White House, despite its crucial role in the AI sector.

The ruling on Anthropic contrasts with a separate federal court decision in California that found a broader ban imposed by the Trump administration against Anthropic to be illegal. The Pentagon decision, based on public procurement regulations, remains in effect. Anthropic has expressed disagreement with the court's decision, stating that the case will be reviewed further.