OpenAI releases framework to track model misalignment

OpenAI released a framework to track, investigate and disclose AI model-misalignment incidents, alongside six reports covering unexpected or concerning model behavior observed over the prior six months. The initiative signals greater transparency and governance around AI safety risks, but contains no financial metrics or immediate commercial impact.