NAIC Big Data Working Group Continues Building a Regulatory Structure

During the NAIC Spring National Meeting, the Big Data and Artificial Intelligence (H) Working Group reviewed its blueprint to build an overarching regulatory edifice to oversee insurers’ use of artificial intelligence systems. The group solicited input from both regulators and stakeholders on its blueprint.

Four foundational pillars provide support for the regulatory structure.

• Define Principles and Assess AI Use. This pillar is nearing completion. The NAIC adopted its principles for artificial intelligence in 2020 and has since conducted surveys of private passenger auto, homeowners, life, and health insurance companies to assess AI usage. Survey results revealed that about 90% of health and auto insurers are likely to use, plan to use, or are exploring the use of AI, compared to only 58% of life insurers. From post-survey discussions, insurers indicated a need for more guidance on regulatory expectations related to governance, AI testing, and oversight of their AI programs.
• Develop AI Risk Evaluation Tools. The main vertical support for this pillar is the NAIC’s model bulletin on the use of AI systems by insurers, which sets forth expectations for the governance, development, acquisition, and use of AI. To further reinforce this pillar, the working group seeks to develop standardized data collection methods and regulatory tools to identify and assess financial and market risks associated with AI use. Additionally, the group is working on an insurer self-audit questionnaire to provide guidance on meeting regulator expectations.
• Regulatory Oversight and Accountability. While construction of this pillar has not yet begun, the working group is “pouring the footers” based on governance, transparency, and accountability. Once completed, this pillar will uphold regulator expectations on matters such as including guidance on when a human should be in the loop and reliance on AI alone would not be permitted. It would also promote the types of disclosures that should be made to stakeholders regarding the use of AI, data use, reasons for decisions, and adverse consumer outcomes. In building this pillar, the working group will invite other NAIC trades to collaborate.
• Identify and Address Gaps in AI Evaluation. This pillar looks toward the future, as it aims to support the identification of emerging AI risks and the development of solutions to address them. Continuing the collaborative effort across the NAIC, this pillar seeks to provide regulators with a framework to identify potential issues before they arise, sealing any gaps between this pillar and the current NAIC structure.

During its meeting, the Innovation, Cybersecurity, and Technology (H) Committee emphasized the importance of the second pillar — developing AI risk evaluation tools — and the need for regulatory evaluation tools and the self-audit questionnaire. The committee sees the self-audit questionnaire as a level for insurers to ensure that their use of AI aligns with regulator expectations and identified it is a key priority for the working group.

All told, the AI regulation project continues to be built out. The working group has begun surveying the land, preparing tools to assess risks and set forth regulator expectations. All the while, other NAIC trades continue to be brought together, helping to support the pillars for a strong regulatory structure. The group’s blueprint is detailed, which will help to keep construction on track.

This article was co-authored by Carlton Fields law clerk Jake Heiges.