NIST Launches AI Model Evaluation Program to Benchmark Performance on Blind Test Data - PYMNTS.com

The U.S. government agency NIST is launching a program to independently test and evaluate AI systems on standardized benchmarks. This matters because independent evaluation helps ensure AI systems work reliably before being deployed in real-world applications.