At the recently concluded National Association of Insurance Commissioners (NAIC) Summer National Meeting in Chicago, various working groups presented reports on the ongoing integration of artificial intelligence (AI) and machine learning (ML) in the insurance industry, particularly focusing on life insurers. The discussions underscored the complexity and breadth of efforts needed to regulate the rapidly evolving technological landscape.
The Accelerated Underwriting (A) Working Group announced the completion of its "Regulatory Guidance and Considerations." This development, which has been referred to the Market Conduct Examination Guidelines (D) Working Group, aims to inform revisions of the Market Regulation Handbook. The guidance provides a structured framework for regulatory bodies reviewing life insurers' use of accelerated underwriting programmes, consisting of three core areas: regulatory considerations, strategies for review, and requests for information. This guidance is rooted in the NAIC’s adopted principles on AI from 2020 and a model bulletin on AI usage by insurers, approved in 2023.
The Accelerated Underwriting Working Group's referral highlighted the need for specific guidelines within the Market Regulation Handbook to alert market conduct examiners about the novel data and processes that life insurers employ in accelerated underwriting. Subsequently, the Life Insurance and Annuities Committee approved both the regulatory guidance and the referral.
Additionally, the Life Workstream of the Special (EX) Committee on Race and Insurance revealed progress on a forthcoming survey designed to query life insurers regarding their utilisation of criminal history in underwriting processes. This survey is part of a broader effort to address potential biases and ensure fairness in underwriting decisions, with plans to convene in late September to deliberate further on the survey.
The establishment of the Third-Party Data and Models (H) Task Force marks another significant step by the NAIC. This task force is charged with addressing regulatory concerns over insurers’ use of third-party data and models, critical components in the modern insurance landscape. It is currently assessing the applicability of several regulatory frameworks, including "risk-focused," "market analysis," and the "Colorado" approach, to ensure robust governance.
Furthermore, the Big Data and Artificial Intelligence (H) Working Group has concluded comprehensive surveys on the use of AI and ML by personal auto, homeowners, and life insurers and is now initiating a similar survey focused on health insurers. They are also responding to consumer group requests for follow-up on former surveys, starting with target meetings with auto insurers.
Another notable discussion involved the exploration of potential racial bias in AI and ML models through the lens of Bayesian Improved First Name and Surname Geocoding (BIFSG). A presentation highlighted inherent limitations in this method, exemplified by its erroneous inference of racial identity during a demonstration.
Throughout the meetings, Iowa Insurance Commissioner Doug Ommen emphasised the necessity for enhanced collaboration among different NAIC groups to efficiently manage the numerous initiatives currently underway. The collective goal remains to ensure that all regulatory frameworks are thorough, up-to-date, and effective in dealing with the dynamic technological changes within the industry.
Source: Noah Wire Services