US AI Safety Institute Teams Up with OpenAI & Anthropic to Test New Models Before Public Release
The U.S. AI Safety Institute, working closely with its U.K. counterpart, will provide feedback to Anthropic and OpenAI on how to improve the safety of their models.



