In an ultimatum issued by the Pentagon less than 24 hours before the deadline, Anthropic has rejected the Defense Department’s demands for unrestricted access to its AI.
It is the culmination of a dramatic exchange of public statements, social media posts and behind-the-scenes negotiations over Defense Secretary Pete Hegseth’s desire to renegotiate all AI labs’ existing contracts with the military. But Anthropic has so far refused to back down from its two existing red lines: no mass surveillance of Americans, and no lethal autonomous weapons (or weapons licensed to kill targets without any human oversight). OpenAI and
In a statement late Thursday, Amodei wrote, “I believe deeply in the existential importance of using AI to defend the United States and other democracies and to defeat our autocratic adversaries. That’s why Anthropic has actively worked to deploy our models across the War Department and the intelligence community.”
He said the company “has never objected to particular military operations nor attempted to limit the use of our technology in an ad hoc manner” but that “in a narrow set of cases, we believe AI could undermine, rather than defend, democratic values” – specifically mentioning mass domestic surveillance and fully autonomous weapons. (Amodei noted that “partially autonomous weapons … are critical to defending democracy” and that fully autonomous weapons could eventually “prove critical to our national defense”, but “today, frontier AI systems are not reliable enough to power fully autonomous weapons.” He did not rule out approving Anthropic for future use of fully autonomous weapons by the military, but noted that they were not ready yet.)
The Pentagon has reportedly already asked major defense contractors to assess their reliance on Anthropic’s cloud, which could be seen as a first step toward designating the company a “supply chain risk” – a public threat the Pentagon only recently created (and a classification typically reserved for threats to national security). The Pentagon was also reportedly considering invoking the Defense Production Act to enforce Anthropic’s compliance.
Amodei wrote in his statement that the Pentagon’s threats do not change our position: We cannot in good conscience accept their request. He also wrote that “If the Department wishes to offboard Anthropic, we will work to enable a smooth transition to another provider to avoid any disruption to ongoing military planning, operations or other critical missions. Our models will remain available as long as needed on the detailed terms we have offered.”
<a href