Anthropic is letting METR researchers look under the hood with employee-level access

Anthropic is granting employee-level access to its systems to the Model Evaluation and Transparency Research (METR) team, allowing them to publish their own assessments of the company's risks, incidents, and practices. This move comes with Anthropic giving up its right to block any publication by METR. This is a significant development in the field of AI transparency and accountability.