Insurance AI Gateway NAIC | Moring
The NAIC shapes what an insurance AI gateway must demonstrate: a written AI program, governance and records for AI-supported decisions, and controls that work in practice. State adoption of the model bulletin plus the 2026 regulatory agenda make gateway-level enforcement and per-decision evidence the practical way carriers answer examination questions.
Short answer
The NAIC shapes what an insurance AI gateway must demonstrate: a written AI program, governance and records for AI-supported decisions, and controls that work in practice. State adoption of the model bulletin plus the 2026 regulatory agenda make gateway-level enforcement and per-decision evidence the practical way carriers answer examination questions.
How it works
The National Association of Insurance Commissioners has moved AI from principles to supervisory expectations, and those expectations land on carriers as concrete control requirements. AI is now used across the insurance industry for underwriting, claims, fraud detection, and servicing, and the NAIC's work on artificial intelligence, big data, and model governance sets the frame in which those uses must be documented and defended. An AI gateway is where most of that documentation becomes automatic.
The centerpiece is the model bulletin on the use of artificial intelligence systems by insurers, adopted in December 2023 and adopted by states since. It expects insurers to maintain a written AI systems program covering governance, risk management, and internal controls, with controls proportionate to the risk a system poses to consumers. Nearly half of US states have adopted it, which means the expectations are examination-ready in most jurisdictions where a carrier writes business.
The NAIC's 2026 regulatory agenda extended the picture. Commentary on the agenda notes two new compliance obligations that the earlier Insurance Data Security Model Law did not meaningfully address, including AI governance itself. Carriers that spent 2024 and 2025 building information security programs and vendor controls found that the AI layer adds requirements their existing program does not satisfy, particularly around runtime enforcement and decision records.
At the NAIC Summer National Meeting in August 2026 in Columbus, Ohio, regulators sharpened their focus on AI, third-party data and models, cybersecurity, and insurance privacy. Three themes emerged: regulators are supplementing principles-based guidance with supervisory tools and examination-readiness efforts, they are focusing on whether governance controls work in practice, and they are reinforcing that insurers remain responsible for the technology they deploy, including third-party components.
Those themes describe exactly what a gateway demonstrates. Examination readiness is easier when every model call carries an identity, a policy version, and an outcome in a record the carrier owns. Controls working in practice is precisely what an inline enforcement layer shows, in contrast to policy documents that describe controls nobody operates. And responsibility for third-party technology is answered with gateway records of what vendor models actually produced, under which version, for which customers.
For carriers planning gateway investment, the NAIC trajectory suggests durability. Supervisory tools and examination readiness are being built now, and insurers remain responsible for outcomes regardless of which vendor supplied the model. A gateway that enforces the written program at the request layer converts an evolving regulatory landscape from a compliance scramble into an operating property of the carrier's AI traffic.
Frequently asked questions
What does the NAIC expect insurers to demonstrate about their AI systems?
A written AI systems program covering governance, risk management, and internal controls, with controls proportionate to the risk each system poses to consumers. Decisions made or supported by AI must satisfy existing insurance law across underwriting, rating, claims, marketing, and fraud detection. An AI gateway supplies the operational evidence: identity, policy version, and outcome recorded on every model call the carrier makes.
How has NAIC activity changed for insurers in 2026?
The regulatory agenda added AI governance obligations that the earlier Insurance Data Security Model Law did not address, and the Summer National Meeting in August 2026 emphasized supervisory tools, examination readiness, and continued insurer responsibility for third-party technology. Regulators are shifting from principles-based guidance toward testing whether controls actually operate, which favors carriers whose enforcement produces records automatically.
Why does examination readiness favor an AI gateway approach?
Examiners increasingly ask whether governance controls work in practice rather than whether documents exist. A gateway answers with operating evidence: every call carries an attributable identity, the policy version applied, and the recorded outcome. A carrier can retrieve a specific decision's record on request instead of reconstructing behavior from memory and vendor dashboards, which is the difference between demonstrating and asserting.
Does NAIC influence apply to carriers in states that have not adopted the bulletin?
Substantially, yes. The model bulletin functions as a template adopted by a growing majority of jurisdictions, and the NAIC continues to build supervisory tools and examination practices that diffuse across states. Carriers operating multistate also face their own actuarial and market-conduct standards. Building gateway enforcement once, against the strictest adopted expectations, is more economical than retrofitting controls state by state as adoption spreads.