NAIC AI Supplement Moves From Principles to Exam Questions

Insurance regulators are turning artificial intelligence principles into examination questions. At its Summer National Meeting on Thursday, the National Association of Insurance Commissioners updated its AI Risk Evaluation Supplement, a set of inquiries that state examiners can use during market conduct reviews, financial examinations or stand-alone AI inquiries.

The working group renamed the document from "AI Systems Evaluation Tool" to avoid confusion. It is not a certification, a rating or a new law. The supplement gives regulators a common structure to collect evidence on how an insurer builds, validates, monitors and governs AI systems.

The timeline has concrete milestones. The pilot includes California, Colorado, Connecticut, Florida, Iowa, Louisiana, Maryland, Pennsylvania, Rhode Island, Vermont, Virginia and Wisconsin. Participating states have used the supplement in different ways: some embedded it in scheduled exams, others treated it as an ad hoc questionnaire. The NAIC cautioned that every state may not finish by Sept. 30. Feedback from the pilot will feed version 5.0, which gets a 30-day public exposure period in September. Version 6.0 will have a 14-day public exposure afterward. Regulators expect to consider version 7.0 for adoption at the fall national meeting.

The supplement's scope shows where insurers will face the most scrutiny. Regulators want to know a system's purpose, data sources, training data, validation procedures and risk classification. They are also asking for documentation, performance monitoring, change histories and evidence that a company can audit and explain an AI-generated outcome.