A decade ago, "sustainable" stopped being a marketing word and started being a disclosure obligation. Companies that made green claims without the receipts found themselves on the wrong end of regulatory scrutiny, securities suits, and brand damage. AI is following the same arc, and insurance carriers will feel it earlier than most.
The signals are converging fast. In November 2025, the SEC's Division of Examinations released its 2026 Examination Priorities. AI moved from "emerging fintech" to a cross-cutting risk that touches cybersecurity, emerging technology supervision, and the accuracy of registrant representations. The Division said it will review registrant representations about AI capabilities for accuracy and assess whether firms have adequate policies to monitor and supervise their use of AI. Industry counsel has consistently read this as the SEC's "AI washing" priority.
A March 2026 WTW analysis found that almost all AI-related securities class actions filed to date involve some form of AI washing. According to Stanford's tracker, shareholders filed 53 AI-related class actions through the first half of 2025 alone.
For carriers, the SEC line is only the start. Insurance has its own AI audience, its own evidence standards, and the audit infrastructure is being built right now.
Most companies that talk publicly about AI face one regulator. Insurance carriers face several.
A public carrier's investor materials sit under SEC disclosure review and the broader enforcement focus on AI washing. The same carrier's rate filings, market conduct exam responses, and consumer disclosures sit under state insurance laws and the NAIC framework. Producers and reinsurers ask their own questions. Agents repeat carrier claims to consumers. A single description of an AI underwriting model can be tested in five venues with five different evidentiary standards. Few of those venues talk to each other.
This is the structural reason AI washing in insurance is not just an extension of the securities concept. State insurance departments have been auditing models for solvency, rate adequacy, and unfair discrimination for decades. AI is a new class of model, but the audit reflex is old. Examiners know how to read a documentation package and decide whether the controls match the claim.
The NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers has now been adopted in roughly 24 to 25 states. Adopting states expect insurers to maintain a written AI program, govern AI risk across the lifecycle, document model development and testing, and oversee third-party AI vendors.
The next step is operationalization. The NAIC's AI Systems Evaluation Tool is in a multistate pilot, with twelve states participating. Formal adoption is anticipated at the 2026 Fall National Meeting. The Tool is designed to give examiners a standardized way to walk through a carrier's AI governance program during market conduct and financial examinations.
When that Tool goes general, examiners across the country will share the same checklist. A carrier whose written program does not match what is in production will not get a courtesy heads-up. The mismatch will surface during an exam.
Most AI in insurance is not built by the carrier. It is licensed from a vendor, embedded in a core platform, or delivered through a third-party administrator. When a carrier makes a public claim about AI, the substance behind that claim almost always lives somewhere else.
This is where AI washing risk concentrates without anyone noticing. A vendor's marketing copy describes a model in aspirational terms. The carrier procures the vendor and reuses the same language in its own materials. The documentation needed to substantiate the claim sits with the vendor, where it cannot be audited on the carrier's timeline. When a regulator or plaintiff asks the carrier to show its work, the carrier is responsible for the answer regardless of who built the system.
A defensible AI program treats vendor AI as the carrier's AI for documentation purposes. Contracts, model cards, validation reports, monitoring evidence, and incident records have to be obtainable on demand. The carriers that get this right early carry fewer surprises into renewal and exam cycles.
The cases regulators have brought so far are the obvious ones. The cases that will define the next two years are subtler. Examples worth stress-testing inside any carrier:
None of these are fraud. All of them create the documentation gap that AI washing claims are built on. The pattern is consistent enough that there is a name for it now: an AI governance problem sitting inside an otherwise functional company.
The reframe is simple. AI washing is not really about AI. It is about attestation. Every public claim is an implicit attestation, whether or not anyone calls it that. The question is whether the governance program can keep pace with the volume of attestations the rest of the company is creating.
Five questions worth running through this quarter:
A carrier that can answer these is not insulated from regulatory action. It is positioned to handle it.
The lesson from a year of AI washing actions is straightforward. AI policy alone does not protect a company. A policy says what should happen. Controls show that it did. The space between those two is where an AI governance failure becomes an AI washing claim, and where the AI governance auditing record decides which way the story goes.
Monitaur was built for this. Our platform helps insurance risk, compliance, and data science teams move from policy to proof: documenting the models in production, validating that controls are operating as designed, monitoring vendor AI alongside in-house AI, and producing the evidence that supports every external claim. When a regulator, an auditor, a reinsurer, or a plaintiff asks how a carrier knows its AI does what it says, the answer is already in hand.
Regulators are paying attention. The carriers that get ahead of this will not be the loudest about AI. They will be the ones who can prove it.
Want to learn more on how to evaluate your maturity for the NAIC tool? Sign up to reserve your spot in our upcoming webinar.
