Our framework for reporting model misalignment - OpenAI
As AI systems grow ... This includes new ways for models to act without authorization, coordinate with other models, or evade oversight; failures ...
- Topics
- Governance & Risk ManagementAI Safety & Existential Risk
- Mentions
- OpenAI
- Tracked via
- “AI Falures”