Federal Court Backs Pentagon, Rules Anthropic’s AI Safety Guardrails Are a 'Supply Chain Risk'

The U.S. Court of Appeals for the D.C. Circuit ruled 2-1 to allow the Pentagon to formally designate Anthropic as a supply chain risk. The legal battle stems from a February dispute where President Trump and Defense Secretary Pete Hegseth accused the AI lab of endangering national security. The court majority stated that Anthropic's hardcoded safety restrictions—designed by the lab to prevent mass surveillance and autonomous weapons—constitute a structural risk because the military might require the model to perform those exact functions. The Defense Department has already started ripping Claude out of its systems, and federal contractors are now explicitly prohibited from using Anthropic products for military tasks. The ban only applies to defense operations, leaving Anthropic free to operate across the rest of the federal government, while a separate California court previously found the Pentagon's parallel retaliation against the lab unlawful. “The Department reasonably feared that Anthropic might manipulate Claude’s design to prevent it from performing national-security functions that the Department deems contractually authorized and necessary.” — Judge Gregory G. Katsas