On February 24, 2026, U.S. Secretary of Defense Pete Hegseth held a meeting with Dario Amodei, the CEO of artificial intelligence firm Anthropic. During the meeting, Hegseth demanded that Amodei sign documents by Friday, which would grant the U.S. military unrestricted access to Anthropic's AI technology. He warned that failure to comply would result in penalties under the Defense Production Act. Hegseth cited Boeing as a precedent, stressing that there should be no limitations on the military's use of AI products.

However, Anthropic stood firm in its refusal to allow its AI to be employed for mass surveillance of U.S. citizens or for autonomous decision-making regarding physical attack targets. Amodei reiterated these two non-negotiable principles, voicing concerns that the insufficient reliability of AI could potentially lead to unintended escalations or mission failures. Previously, the U.S. military had already utilized Anthropic's AI model, Claude, in operations conducted in Venezuela, a move that had sparked internal concerns within the company.

Currently, both sides are engaged in negotiations over the usage policies. Anthropic is striving to ensure that its technology is not misused, while the Pentagon is pushing for AI companies to allow the military to utilize their models in any scenario that is legally permissible.