The Department of War has demanded that Anthropic’s AI models be made available to it for "any lawful use." Anthropic has, however, refused to cross its strict ethical red lines — no fully autonomous weapons and no mass domestic surveillance.