We updated our Preparedness Framework for tracking & preparing for advanced AI capabilities that could lead to severe harm. The update clarifies how we track new risks & what it means to build safeguards that sufficiently minimize those risks. https://openai.com/index/updating-our-preparedness-framework/