What ChatGPT says about its own use: AI can legitimately assist humans by detecting incoming weapons, navigating, gathering intelligence or highlighting possible threats. But the irrevocable decision to intentionally kill a person or launch nuclear weapons should never be delegated to an algorithm.
