OpenAI Launches Model Misalignment Reporting Framework

OpenAI introduced a structured framework to track, investigate, and publicly disclose instances of model misalignment across the AI lifecycle. The process begins when an employee flags an issue, leading to technical investigation and sorting into three review tracks for disclosure.