The Architecture of Refusal — Military AI and AI Guardrails

gerl.dev

The Pentagon designated Anthropic a security threat over two AI guardrails. What 'any lawful purpose' means depends entirely on who defines 'lawful.'