Developing: Pentagon Gives Anthropic Friday Deadline to Drop AI Safety Guardrails — Or Face Blacklisting and Defense Production Act
Defense Secretary Pete Hegseth delivered an ultimatum to Anthropic CEO Dario Amodei on Tuesday: grant the U.S. military unrestricted access to the company's artificial intelligence models by Friday evening, or face severe consequences including potential blacklisting from all government contracts and invocation of the Defense Production Act. The confrontation, which took place during a meeting at the Pentagon, marks the most dramatic escalation yet in a growing rift between the Trump administration and one of America's leading AI companies over the ethical boundaries of military AI deployment. At the heart of the dispute is Anthropic's insistence on maintaining two red lines: its AI systems should not be used for fully autonomous lethal targeting decisions without human oversight, and they should not be deployed for mass surveillance of American citizens. The Pentagon, which rebranded itself the Department of War under the current administration, has demanded that Anthropic agree to "all lawful use cases" without any company-imposed limitations — a framing that Anthropic's leadership views as dangerously open-ended. The standoff has thrust questions about AI ethics, corporate responsibility, and military power into the center of a high-stakes policy showdown with no clear precedent. The clash carries enormous implications not just for Anthropic, which holds a $200 million defense contract and was the first AI company cleared for classified military networks, but for the entire AI industry. How this dispute resolves could set the template for the relationship between Silicon Valley and the Pentagon for decades to come — determining whether AI companies retain any say over how their technologies are used in warfare and intelligence operations.