Pete Hegseth wants unfettered access to Anthropic’s models for the military

In a significant escalation between the U.S. Department of Defense and the AI company Anthropic, Defense Secretary Pete Hegseth has issued an ultimatum for the company to grant access to its technology for military applications. Hegseth warned that Anthropic could be removed from the department's supply chain if it does not agree to unrestricted use of its models by the end of the week. This confrontation stems from Anthropic's reluctance to provide unfiltered access to its AI systems for sensitive military operations, including those that involve domestic surveillance and potentially autonomous combat missions without human oversight. Hegseth's meeting with Anthropic's CEO, Dario Amodei, in Washington, was marked by high tension as the defense chief outlined the severe repercussions of non-compliance. According to sources privy to the discussions, Hegseth emphasized that the deadline for Anthropic to comply was set for Friday at 5:01 PM. If the company fails to meet these demands, the Secretary indicated he would resort to the Defense Production Act—a Cold War-era law that allows the government to control domestic production for national security purposes. This move could compel Anthropic to comply with military requirements against its will. The Pentagon plans to categorize Anthropic as a “supply chain risk” should it refuse to cooperate. An official stated, “You can’t lead tactical operations by exception,” while asserting that the discussions do not pertain to issues of mass surveillance or the deployment of autonomous weapons. In response to the situation, Anthropic has stated that it remains engaged in constructive discussions regarding its usage policies, aiming to align its operations with the government's national security objectives while ensuring that its models are employed responsibly and reliably.