Most conversations about AI governance still treat large language models (LLMs) as if they are simply chatbots—tools[…]