Book a demo
Topics · AI-CONTROL-PLANE

Insurance AI Control Plane NAIC Bulletin | Moring

The NAIC Model Bulletin, adopted December 4, 2023 and now adopted by about half of US jurisdictions, requires insurers to maintain a written AIS Program with governance, risk management, and internal controls across the AI lifecycle. A control plane turns those requirements into runtime enforcement and the record an examiner asks to see.

On this page
  1. Short answer
  2. How it works
  3. FAQs
  4. Related reading
  5. Sources

Short answer

The NAIC Model Bulletin, adopted December 4, 2023 and now adopted by about half of US jurisdictions, requires insurers to maintain a written AIS Program with governance, risk management, and internal controls across the AI lifecycle. A control plane turns those requirements into runtime enforcement and the record an examiner asks to see.

How it works

The NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers was adopted on December 4, 2023, as a template for state insurance departments. It is explicit that it establishes no new substantive requirements; instead it signals how regulators will use existing unfair trade practices statutes and market conduct examination authority to review how insurers develop, acquire, and use AI systems that make or support consumer-impacting decisions. As of April 2026, 25 jurisdictions had adopted it, with four more maintaining insurance-specific guidance of their own.

The bulletin defines an Artificial Intelligence System broadly: a machine-based system that, for given objectives, generates outputs such as predictions, recommendations, or decisions that influence insurance-related decisions. The definition captures machine learning models, generative AI, and agentic systems across underwriting, rating, claims handling, marketing, and fraud detection. The practical effect is that any automated tool touching a consumer-facing insurance outcome belongs in the insurer's AI inventory, regardless of whether the owning business unit calls it AI.

Three connected requirements form the core. Governance: oversight across the full lifecycle, from development or acquisition through validation, deployment, and monitoring, with accountability at board and senior management level. Risk management: testing for accuracy, reliability, and the potential for unfair discrimination, with the methodology left to the insurer provided the risks are actively assessed. Internal controls: the policies and control points that tie the first two together operationally, ensuring the program runs as designed rather than as documented.

Section 4 defines the examination posture. 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, and those questions apply regardless of the existence or scope of a written AIS Program. The bulletin also places accountability for third-party AI squarely on the carrier.

This is where the control plane becomes the implementation vehicle. A governance committee's approvals, a risk-tiering methodology, and a vendor oversight workflow are organizational artifacts; the bulletin's language implies examiners look for proof the policy is operational, a person or committee responsible, a record of decisions, and controls that trace to specific systems in production. The control plane supplies exactly that: runtime enforcement of the controls the program describes, plus a sealed action record tying each consequential decision to its policy, its approver, and its outcome.

Implementation guidance follows from the structure. Maintain the AIS Program as a living artifact reviewed on a cadence, not a launch document. Map every production AI system, including agents, to the program's risk tiers. Wire the internal controls into the request path so enforcement is real. And rehearse the examination: for a sampled decision, produce the complete evidence chain from the control plane's record in the time an examiner would expect.

Frequently asked questions

What is the NAIC Model Bulletin and when was it adopted?

The Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, adopted by the NAIC on December 4, 2023, as a template for state insurance departments. It creates no new substantive law; it signals that existing unfair trade practices and unfair discrimination rules apply fully to AI-supported decisions, and sets expectations for governance, documentation, and examination. By April 2026, 25 jurisdictions had adopted it, with four more carrying their own insurance-specific AI guidance, so carriers should check the status of every state where they write business.

What does the bulletin require in an insurer's AIS Program?

Three connected elements: governance, meaning lifecycle oversight with board and senior management accountability; risk management, meaning testing for accuracy, reliability, and potential unfair discrimination, with methodology left to the insurer; and internal controls, the policies and control points that make the program operational rather than documented. The definition of an AI system is broad, capturing predictive models, generative AI, and agentic systems across underwriting, rating, claims, marketing, and fraud detection, so the program's inventory must reach every system that touches a consumer-facing insurance outcome.

What questions will an examiner ask under the bulletin's Section 4?

How the insurer develops, deploys, and uses its AI systems, including any specific model or application; the outcomes of those systems, including adverse consumer outcomes; and the governance framework, risk management, and internal controls behind them. These questions apply regardless of whether a written AIS Program exists, so a carrier without one is still expected to answer. The defensible answers are retrievable records: the action trail for a specific decision, the approval that governed it, and the monitoring and testing evidence for the system that produced it.

How does a control plane satisfy the bulletin's internal controls expectation?

By making the controls operational. Examiners look for evidence that policy is running, not filed: a person or committee accountable, a record of decisions made under the policy, and controls that trace to specific systems in production. The control plane enforces the program's boundaries in the request path, holds consequential actions for named reviewers, and seals each action record with its policy version and approval. The result is that the AIS Program and the running system agree, which is precisely the consistency the bulletin's examination posture is designed to test.

Sources

  1. content.naic.org
  2. trussed.ai
  3. vijil.ai
  4. moring.ai
  5. thedataexperts.us
  6. larsco.com
  7. plantemoran.com
  8. trussed.ai
  9. - artificialintelligenceact.eu
  10. - eur-lex.europa.eu