White paper · PDF · Version 1.0, August 2026
Independent AI Evaluation Assurance
A framework for determining whether AI performance claims can be trusted. It sets out seven evaluation dimensions, the E0–E4 evidence hierarchy, Evaluation Confidence Levels, the Production Evaluation Gap between laboratory and operational capability, and the Maximum Credible Evaluation Failure.
Download the white paper About the assurance service