Insurtech Advocates Warn State Regulators: Vague Third-Party Data Definitions Could Harm AI Innovation and Consumers

State Regulators Urged to Clarify Third-Party Data Definitions for AI in Insurance

State insurance regulators are moving to define key elements of AI regulation, particularly around third-party data and models. The National Association of Insurance Commissioners (NAIC) Third-Party Data and Models Working Group, during its May 22 meeting, agreed on the need to establish definitions to support AI models that are more dynamic than traditional predictive or deterministic models.

Most state insurance regulators have experience with traditional models, not the evolving AI-driven ones. Gennady Stolyarov, lead actuary at the Nevada Division of Insurance, pointed out this regulatory gap, highlighting the necessity for clear definitions to guide oversight.

Concerns from the American InsurTech Council

The American InsurTech Council (AITC), which advocates for ethical and tech-driven innovation in insurance, has criticized the NAIC working group's approach. Scott Harrison, AITC co-founder, warns that vague or imprecise definitions risk creating regulatory inconsistencies between states. This could lead to confusion over which entities or use cases fall under regulation, resulting in disputes and uncertainty in the marketplace.

Harrison explained that such uncertainty may:

• Hinder innovation
• Increase compliance costs
• Create regulatory gaps potentially harmful to consumers
• Distract from developing focused regulatory standards

Questions Over the Purpose of Defining Third-Party Data

Jason Lapham, deputy commissioner for property and casualty insurance at Colorado's Division of Insurance and upcoming director of the NAIC working group, has not clearly stated the purpose behind setting strict definitions. Harrison, in a June 23 letter to Lapham, expressed concern that rigid definitions could stifle innovation and complicate compliance.

He emphasized that the “dynamic, rapidly evolving and innovative nature” of third-party services to insurers may not fit into strict categories. Rigid definitions could lead to confusion in contracting, impose unnecessary regulatory burdens, and cause inconsistent interpretations across states.

Governance and Accountability in AI Use

Harrison advocates for insurers to implement governance structures for AI use rather than rely solely on prescriptive definitions. The NAIC has already introduced a Model AI Bulletin, providing guidance on AI regulation that 24 states have adopted.

During the May 22 meeting, Birny Birnbaum, director of the Center for Economic Justice, stressed the importance of establishing definitions around third-party data and vendors, particularly distinguishing them by data life cycle and data type. Meanwhile, the NAIC's Big Data and Artificial Intelligence Working Group raised the critical question of responsibility for the accuracy of third-party data.

Shannen Logue, deputy commissioner of product regulation at Pennsylvania Insurance Department, underlined that accountability and transparency are essential when insurers work with third-party data providers.

Implications for Insurance Professionals

For insurance professionals, this ongoing debate highlights the need to stay informed about regulatory developments regarding AI and third-party data. Clear governance practices and transparency in vendor relationships will be crucial as AI tools become more integrated into underwriting, claims, and risk assessment workflows.

Those interested in strengthening their AI knowledge to better navigate these challenges may find value in specialized courses on AI applications in insurance.