Australia wants time-bound reporting after an AI agent crossed the boundary
Australia's response to an OpenAI agent breach turns incident disclosure into an architecture problem: preserve evidence, report preliminary facts early, and keep human decisions explicit.