Maryland Insurance Commissioner Kathleen A. Birrane has issued industry Bulletin 24-11: The Use of Artificial Intelligence Systems in Insurance, which provides regulatory direction to insurance companies regarding their use of artificial intelligence.
The AI bulletin is premised on the model adopted unanimously by the National Association of Insurance Commissioners in December 2023 under the leadership of Commissioner Birrane, who serves as chair of the association’s Innovation, Cybersecurity and Technology Committee. The model bulletin was the product of a collaborative bi-partisan effort by U.S. regulators “to establish clear, principals-based guidelines that promote the responsible and ethical use of AI systems by the insurance sector,” Commissioner Birrane said.
The bulletin reminds carriers that when they use advanced analytical and computational technologies like AI to make decisions that impact consumers, they must comply with all applicable insurance laws and regulations, including Maryland’s unfair trade practice and anti-discrimination laws. The bulletin sets out the Insurance Administration’s expectations for the development and deployment of AI systems, including the establishment of a written program that focuses on governance, risk management and internal controls. Because carriers are responsible for any unfair discrimination that results from their use of predictive models and other AI systems, the bulletin encourages carriers to use validation and testing methods to identify error and unfair bias. The bulletin also advises insurers of documentation that the Insurance Administration is likely to request during an investigation or examination.
“AI technologies have become, and will continue to be, a critical tool for closing the insurance gap,” Commissioner Birrane said. “But the very nature of the technology creates risks that require strong controls, particularly where the technology is deployed to make decisions that can harm consumers.”
To that end, the bulletin:
- Sets the expectation that carriers will establish a written governance, risk management and controls for its use of AI systems that address the entire life cycle of both the AI system and the relevant insurance product and will be prepared to provide that written program and evidence of compliance with the program in response to inquiries by the Insurance Administration about AI system use.
- Identifies the elements that a carrier should consider in developing its AI systems program, including internal oversight and accountability oversight, data standards and controls, validation and testing, and diligence on third-party data and models.
- Directs that controls utilized with respect to any AI use case be commensurate with the risk of consumer harm that could result from that use.
”Our expectation is that carriers will use the principles and the guidelines set out in the bulletin to establish an AI program that is designed to ensure that its use of AI is ethical, responsible and compliant with regulatory requirements,” Commissioner Birrane said.
