OpenAI Strengthens Transparency with Expanded Incident Disclosure Framework OpenAI announces new safety protocols to track and disclose AI model misalignment incidents, strengthening transparency... 17 Sep 2026