OpenAI finds risks in fine-tuning AI on bad data - Tech in Asia
AI Safety Auditing: Creating methods to assess and identify misalignment risks in training data. Fine-tuning Protocols: Developing guidelines to ...
Read the original at Tech in Asia ↗
- Topics
- AI Safety & Existential RiskGovernance & Risk Management
- Mentions
- OpenAI
- Tracked via
- “AI Risks”