GPT-5.1 Ethics Framework: The CAN/CANNOT Binary as Innovation

GPT-5.1's ethics checklist introduces a methodological innovation: the CAN/CANNOT binary at the top of each template • explicitly enumerating what evidence types are permissible and what proxy inferences are forbidden — creating a guardrail structure that is both clear and enforceable