Rogue Behavior: OpenAI Reveals More Model Misalignment Incidents

Dark Reading · 2 hours ago General

The AI giant disclosed six examples of concerning model activity and published a new framework for investigating and disclosing such incidents.