The v1.0 AILuminate benchmark from the MLCommons AI Risk & Reliability working group is the first AI risk assessment benchmark developed with broad involvement from leading AI companies, academia, and civil society. It is a significant step towards standard risk assessment for “chat” style systems.
