At its August 13 meeting, the NAIC's Big Data and AI Working Group said a 12-state pilot of its AI Risk Evaluation Supplement will continue through September. The supplement structures evidence requests for insurer reviews; it is not a certification or new law. Pilot feedback is slated to inform two public exposure drafts before regulators consider adoption at the fall national meeting.
At its August 13 meeting, the National Association of Insurance Commissioners' Big Data and Artificial Intelligence Working Group reported that a 12-state pilot of its AI Risk Evaluation Supplement will continue through September. Official meeting materials show that states are testing the supplement in market-conduct and financial examinations, financial analyses, and more general regulatory inquiries.
The document was previously called the AI Systems Evaluation Tool. The working group renamed it to reduce confusion about its purpose, according to the official materials. PYMNTS reports that the supplement is not a certification program, rating system, or new insurance law; it is a structured way for regulators to request evidence about how insurers use and govern AI.
Twelve states are testing the supplement
The pilot includes California, Colorado, Connecticut, Florida, Iowa, Louisiana, Maryland, Pennsylvania, Rhode Island, Vermont, Virginia, and Wisconsin. Participating regulators are using different processes: some have incorporated the supplement into scheduled examinations, while others are using it as an ad hoc questionnaire.
The NAIC timeline says state input and initial company survey responses will inform version 5.0, which is scheduled for a 30-day public exposure period in September. A version 6.0 is then scheduled for a 14-day exposure before the working group considers version 7.0 for adoption at the fall national meeting. The NAIC cautions that individual state pilots may finish on different schedules and that not all may be complete by September 30.
What examiners may request
The current version of the official supplement asks insurers to quantify AI use by business area and describe governance frameworks, validation and testing, high-risk models, data inputs, monitoring, vendor oversight, consumer impacts, and controls. It also directs regulators to tailor requests to the scope of an examination rather than treat every question as universally required.
For data science and ML teams in insurance, the practical issue is evidence readiness. A team may need to produce a current model inventory, ownership and approval records, validation results, data lineage, monitoring history, and documentation for third-party systems. Those artifacts make it possible to reconstruct how a model was selected, tested, governed, and changed; the pilot is showing how regulators may ask for them through existing supervisory processes. The supplement remains a draft in a pilot-and-exposure process. Its August update is therefore a progress report on a possible common examination framework, not a final nationwide rule.
Key Points #
- 1The NAIC's 12-state AI Risk Evaluation Supplement pilot will continue through September across examinations, financial analyses, and other regulatory inquiries.
- 2Pilot feedback is scheduled to inform a 30-day exposure of version 5.0 and a 14-day exposure of version 6.0 before version 7.0 is considered at the fall national meeting.
- 3The supplement asks for evidence including AI inventories, governance controls, validation, model and data details, monitoring, vendor oversight, and consumer-impact records.
Scoring Rationale #
The 12-state pilot is operationally relevant to insurance ML teams because it translates AI governance expectations into evidence requests used in supervisory reviews. Its impact is meaningful but bounded because the supplement remains in a pilot and public-exposure process rather than being a final nationwide rule.
Sources #
Primary source and supporting public references used for this report.
Practice with real Health & Insurance data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Health & Insurance problems