New York Proposes Guidelines for Fair AI Use in Insurance Underwriting

New York State is taking proactive steps to ensure fairness and transparency in the insurance sector’s use of artificial intelligence (AI) and external data sources. A recent circular letter from Superintendent Adrienne A. Harris of the Department of Financial Services (DFS) highlights the potential risks and responsibilities associated with AI and external consumer data and information sources (ECDIS) in insurance underwriting and pricing.

While acknowledging the benefits of AI and ECDIS in simplifying and potentially enhancing the accuracy of underwriting and pricing, DFS raises concerns about the self-learning nature of AI, which could lead to unfair or unlawful discrimination. This is particularly worrying as it might disproportionately affect vulnerable communities or destabilize the New York insurance market. Another concern is the varying accuracy of external data, which may originate from sources lacking regulation or consumer protections.

Superintendent Harris emphasizes that technological advances in underwriting and pricing should not compromise consumer protection. She insists on DFS’s responsibility to ensure AI in insurance doesn’t perpetuate or amplify systemic biases leading to discrimination. The circular letter seeks feedback on these proposals, with the insurance industry, including the New York Insurance Association, ready to collaborate with DFS to balance innovation and consumer protection.

The proposed guidelines expect insurers to analyze AI and ECDIS for discrimination, demonstrate their actuarial validity, and maintain oversight, transparency, risk management, and internal controls. Insurers must also ensure responsibility for compliance with anti-discrimination laws, irrespective of assurances from third-party vendors.

In December 2023, the National Association of Insurance Commissioners (NAIC) voted to approve a model regulation for AI use in insurance, focusing on governance, documentation, third-party collaborations, and risk management. This move, along with the Financial Stability Oversight Council’s warning about AI’s potential risks to the U.S. financial system, underscores the need for diligent supervision of rapidly advancing AI technologies in the financial sector.