State Regulators Launch Major Review of AI's Role in Auto Insurance Pricing

Regulators Finally Look at How AI Is Setting Your Auto Insurance Rate

State insurance regulators took a significant step in March 2026, formally launching a new working group dedicated to assessing how artificial intelligence, telematics, digital distribution, and large-scale data vendors are reshaping the auto insurance market. The group, appointed at the NAIC's 2026 Spring National Meeting in San Diego on March 25, is tasked with delivering actionable recommendations by the end of 2026.

"Artificial intelligence, new distribution models, national scale vendors, and other technological advances are significantly changing both consumers' expectations and insurers' business models and practices," said Illinois Department of Insurance Director Ann Gillespie, who chairs the NAIC's Market Regulation and Consumer Affairs Committee. "State insurance regulators focus on leading, not just responding."

For most American drivers, the timing matters. AI-based pricing systems are not a future trend — they are already at work today, and most policyholders have no idea what data those systems are using or how to challenge the results.

What AI Actually Does to Your Premium

Traditional actuarial models have always used statistical categories to price risk — age, driving record, vehicle type. What's new in AI-driven underwriting is the scale and opacity of the inputs. Modern machine learning models can simultaneously process thousands of variables, many of which have no obvious connection to how you drive.

Major carriers and their data broker partners — primarily LexisNexis and Verisk — feed these models inputs that can include your credit history, home ownership status, purchasing behavior, geographic movement patterns, and in some cases, connected car telemetry updated in real time. A 2024 investigation revealed that some automakers had been quietly sharing driving behavior data from connected vehicles directly with insurers and their data partners without clear consumer disclosure.

The controversy isn't just about privacy. If an AI model correlates a factor like ZIP code, home ownership, or purchasing habits with accident risk, it can effectively function as a proxy for race or income — even unintentionally. That's the core concern driving regulators to act now.

Which States Have the Strongest Protections Now

Regulation of AI in auto insurance is entirely state-level, and the gap between states is significant. If you live in California or Colorado, you currently have the strongest consumer protections in the country. California's broader regulatory environment — already one of the most active in the country — is examined in our coverage of how new California auto insurance rules are hitting drivers' wallets.

State	Key AI Regulation	Effective Date	What It Does
California	Proposition 103	1988 (ongoing)	Requires actuarial justification for all rating factors; heavily restricts algorithm-only pricing
Colorado	SB 21-169 + SB 24-205	Nov 2023 / Oct 2025 / Feb 2026	Bans unfairly discriminatory algorithm use; requires bias testing and annual compliance reports for auto insurers
23 States + DC	NAIC AI Model Bulletin	Varies by state (through 2025)	Requires written AI governance programs, documentation, and audit procedures
12 States*	NAIC AI Evaluation Tool Pilot	Jan–Sep 2026	Regulators actively reviewing insurer AI systems for discriminatory outputs

*Pilot states: California, Colorado, Connecticut, Florida, Iowa, Louisiana, Maryland, Pennsylvania, Rhode Island, Vermont, Virginia, and Wisconsin. Source: NAIC, 2026. Nationwide rollout expected November 2026.

In states without these protections, insurers have wide latitude to use third-party data models with minimal disclosure requirements. The NAIC working group is expected to address this patchwork by developing model standards — including transparency requirements, consumer disclosure rules when AI drives an adverse decision, and potential audit mandates — that states can adopt uniformly.

What the NAIC Working Group Will Examine

The Market Conduct Regulation Modernization Working Group has a broad mandate that goes beyond AI alone. According to the NAIC's announcement, the group will examine AI underwriting models, telematics data use, digital distribution channels, and the growing role of national-scale vendors selling algorithmic tools to insurers across multiple states.

That last piece — third-party vendor accountability — is particularly significant. Many insurers don't build their own AI models; they purchase them from vendors like LexisNexis Risk Solutions, Verisk, or smaller insurtech firms. Until recently, carriers could largely blame the vendor if a model produced discriminatory results. The NAIC's model bulletin already pushes back on this: carriers are now expected to take full accountability for the AI tools they deploy, regardless of where those tools came from.

The working group will collect broad stakeholder input throughout 2026 and aims to deliver concrete recommendations by year-end. While the NAIC's process moves deliberately — its model laws are advisory, not binding — the standards it sets tend to be adopted by many states in subsequent legislative sessions.

What This Means for Your Auto Insurance Rate

The immediate practical impact of the NAIC announcement is modest — the working group's recommendations won't produce new state laws overnight. But the regulatory environment for AI in auto insurance is clearly tightening, and that has real near-term implications for drivers.

First, if you live in one of the 12 AI Evaluation Tool pilot states, your insurer's pricing model is already under active regulatory scrutiny in 2026. States like Florida, California, Colorado, and Connecticut are part of this pilot — which means insurers operating there are under more pressure to clean up discriminatory or opaque pricing practices now, ahead of the November 2026 nationwide rollout.

Second, under the NAIC Model Bulletin already adopted by more than half of states, carriers are required to provide consumers with timely written explanations when AI influences an adverse underwriting outcome. If your insurer raises your rate significantly or declines to cover you, you have grounds to ask for a written explanation of what factors drove that decision. Texas went even further — its new 2026 law requires insurers to explain policy denials in writing, a model other states are now watching closely.

Third — and this is the gap most consumers don't know about — you already have the right to request your insurance score from data brokers like LexisNexis. That score, built from your credit history and claims data, feeds directly into the AI models pricing your policy. A wrong entry in that score can cost you hundreds of dollars per year.

Looking Ahead: What to Expect by End of 2026

By November 2026, the NAIC's AI Evaluation Tool is expected to expand beyond the current 12-state pilot to all 50 states and D.C. That means regulators in every state will have a standardized framework for auditing insurer AI systems — a significant shift from the current patchwork. The NAIC working group's recommendations will likely accompany this rollout, setting the stage for new model laws in 2027 state legislative sessions.

Consumers in states that have been slower to regulate AI — many in the South and Midwest — may see meaningful new protections within 12-18 months if the NAIC's model standards gain traction. In the meantime, the most important step you can take is to know what data is being used to price your policy and to actively compare rates across carriers, since the AI gap between insurers can easily mean hundreds of dollars per year in premium differences.

Frequently Asked Questions