Insurance companies are either already using or experimenting with AI for both underwriting and claims handling.  And it is certain that the use of AI by the insurance industry will continue to grow.  This poses both possibilities and perils for consumers.  The National Association of Insurance Commissioners has been vigilant in examining these risks, and on December 1, 2023 issued a model bulletin that individual states can adopt to regulate insurance company use of AI.  On February 11, 2025, the New Jersey Department of Banking and Insurance (“NJ DOBI”) adopted the NAIC model bulletin. New Jersey is the twenty-third state to adopt the bulletin (two other states have adopted their own regulatory framework for use of AI in insurance).

The model bulletin, as adopted in NJ DOI Bulletin No. 25-03, issued on February 11, 2025, begins by noting that AI “can present unique risks to consumers, including the potential for inaccuracy, unfair discrimination, data vulnerability, and lack of transparency and explainability.”  As a result, the NJ DOBI requires every insurance company doing business in the state and utilizing “AI Systems” to comply with the applicable enumerated New Jersey legal and regulatory standards, including: (1) Unfair Trade Practices Act; (2) Unfair Claims Settlement Practices Act; (3) Corporate Governance Annual Disclosure Act; (4) Property and Casualty Rating Laws; and (5) Market Conduct Surveillance.

New Jersey licensed insurance companies also are required to “develop, implement, and maintain a written program (‘AIS Program’) for the responsible use of AI Systems that make or support decisions related to insurance practices.  The AIS Program should be designed to mitigate the risk of Adverse Consumer Outcomes, including, at a minimum, the [above] statutory provisions.”

An AI System (AIS) is defined in the bulletin as “a machine-based system that can, for a given set of objectives, generate outputs such as predictions, recommendations, content (such as text, images, videos, or sounds), or other output influencing decisions made in real or virtual environments.  AI Systems are designed to operate with varying levels of autonomy.”

The bulletin stipulates that the controls and processes an insurance company adopts and implements as part of its AIS Program should reflect the company’s assessment of the degree and nature of risk posed to consumers by its AI Systems, considering:

1. the nature of the decisions being made, informed, or supported by AI;
2. the type and degree of potential harm to consumers;
3. the extent of human involvement in final decision making;
4. transparency and explainability of outcomes to consumers; and
5. the extent and scope of reliance on, or use of, third party data, AI Systems, or Predictive Models.

As an overriding principle, the AIS Program should be governed by prevention and mitigation of adverse consumer outcomes and should prioritize transparency, fairness, and accountability.  It should include processes and procedures providing notice to impacted consumers that AI Systems are in use.  The AIS Program also should document the management and oversight of AI Systems, including their development and use, interpretability, and testing.  With respect to Predictive Models, the AIS Program should provide a “narrative description of the model’s intended goals and objectives.”

Insurance companies’ decision-making processes in determining coverage for claims is already opaque, and the increased use of AI Systems surely will create new barriers and challenges to understanding exactly how claim determinations are being made.  This is why it is so critical that the NJ DOBI regulations prioritize accountability and explainability.  They also emphasize the importance of human involvement in AI decision-making.  Members of United HealthCare Medicare Advantage plans already have filed putative class action litigation over the alleged use of AI to deny claims without sufficient human oversight in the medical claims decision-making context.

The NJ DOBI has acted in a timely manner to regulate insurance companies’ use of AI in order to protect consumers.  It is difficult to know what these companies’ AI Systems ultimately will look like and how they will be implemented.  But it is certain that their use will continue to grow, and that both the insurance industry and state regulators will face many questions.  Will insurance companies’ AIS Programs be robust public documentation or will they be exploited as an opportunity for deception?  How will AI-based regulations be enforced?  The NJ DOBI has adopted the NAIC model bulletin to protect policyholders by casting light on the hidden workings of AI within the insurance industry, but consumers need to remain vigilant.