Supposedly, Anthropic refused to give the Pentagon unrestricted access to Claude, its frontier AI model, the only one currently running on classified military networks. They wanted guarantees that there would be zero mass surveillance and no autonomous weapons without a human in the loop, making the final decisions of life or death. The Department of War's message was 'remove those restrictions or lose everything.'
We have long believed that AI should not be used for mass surveillance or autonomous lethal weapons, and that humans should remain in the loop for high-stakes automated decisions. These are our main red lines. It was the morning of February 27. That night, Altman made a deal with the DoW, which did not include any such contractual agreements.