Insurance AI-DLC NAIC Bulletin | Moring
The NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers defines the expectations an insurance AI-DLC must satisfy: a written AIS program, governance across the AI lifecycle, risk management proportionate to use, and documentation an examiner can review. Building the lifecycle around those requirements makes bulletin compliance a by-product of how carrier software ships.
Short answer
The NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers defines the expectations an insurance AI-DLC must satisfy: a written AIS program, governance across the AI lifecycle, risk management proportionate to use, and documentation an examiner can review. Building the lifecycle around those requirements makes bulletin compliance a by-product of how carrier software ships.
How it works
The NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, adopted December 4, 2023 by the association's Executive Committee and Plenary, is the anchor document for AI governance in American insurance. It opens by reminding insurers that decisions or actions impacting consumers made or supported by advanced analytical and computational technologies must comply with all applicable insurance laws, including those addressing unfair trade practices and unfair discrimination.
The bulletin defines an Artificial Intelligence System broadly: a machine-based system that, for a given set of objectives, generates outputs such as predictions, recommendations, or decisions that influence insurance-related decisions. That width deliberately captures machine learning models, generative tools, and predictive algorithms across underwriting, rating, claims, marketing, and fraud detection, whether built internally or licensed. For delivery, it means the systems an AI-DLC produces fall squarely within scope.
Its program requirement is specific. The bulletin directs insurers to develop, implement, and maintain a written AIS program addressing governance across the full AI lifecycle, from development or acquisition through validation, deployment, and ongoing monitoring, with board and senior management oversight; risk management that tests accuracy, reliability, and potential for unfair discrimination; and documentation. Third-party sourcing does not transfer the duty: the insurer remains accountable for vendor-supplied AI, the point carriers most often underestimate.
Section 4 gives the bulletin its operational teeth. In an investigation or market conduct action, an insurer can expect questions about how it develops, deploys, and uses AI systems, the outcomes of those systems including adverse consumer outcomes, and its governance framework, risk management, and internal controls, regardless of the existence or scope of a written AIS program. Documentation produced routinely beats documentation assembled under examination pressure.
An insurance AI-DLC operationalizes every element. The lifecycle's named owners and approval gates instantiate accountability; its evaluation gates supply risk-based testing before release; its monitoring and rollback controls cover deployment and ongoing oversight; and its change records, prompts, approvals, and test evidence constitute the documentation. An insurer shipping under this model can answer Section 4 questions from records it already holds.
Adoption gives the bulletin force in particular jurisdictions, spreading across states from early 2024 onward, with a dozen or more adopting quickly and roughly half of United States jurisdictions on board by 2026. The consequence for carrier engineering is uniform: build the AI-DLC to generate the evidence the bulletin describes, and each adopted jurisdiction's version of it becomes a documentation exercise rather than a redesign.
Frequently asked questions
What is the NAIC Model Bulletin on the Use of AI Systems by Insurers?
Adopted on December 4, 2023, it is the National Association of Insurance Commissioners' model bulletin reminding insurers that existing insurance law applies fully to decisions made or supported by AI systems, and setting expectations for governance, risk management, and documentation. For an insurance AI-DLC, it defines what the delivery lifecycle must prove about the systems it builds: accountability, proportionate risk management, and records an examiner can review.
What is an AIS program under the bulletin?
It is the written program the bulletin expects insurers to develop, implement, and maintain for AI systems that make or support regulated decisions. It covers governance across the AI lifecycle from acquisition through deployment and monitoring, risk-based testing including checks for unfair discrimination, and defined accountability with senior-level oversight. An insurance AI-DLC supplies the operational layer: agent inventories, evaluation gates, owners, and change evidence.
What questions does Section 4 of the bulletin anticipate?
It says an insurer can expect to be asked how it develops, deploys, and uses AI systems, what outcomes those systems produce including adverse consumer outcomes, and how its governance framework and internal controls operate, regardless of whether a written AIS program exists. An insurance AI-DLC answers from accumulated records: per-change provenance, evaluation results, approvals, and monitoring evidence.
Does the bulletin cover vendor-supplied AI and delivery tools?
Yes. Its AIS definition is deliberately broad, capturing machine learning models, generative tools, and predictive algorithms whether built or licensed, and the insurer remains accountable for third-party AI. In an insurance AI-DLC that includes the models inside developer tooling and the vendor components built into carrier systems, so vendor provenance, evaluation coverage, and integration records belong in the program documentation.