Artificial intelligence (AI) is changing how insurance premiums are calculated. As insurers become better at predicting individual risks, premiums are becoming more segmented. However, this trend raises concerns about reduced access to insurance for high-risk groups and potential discrimination. Experts warn that while AI may enhance the efficiency of insurance, it could undermine the fundamental principle of mutual aid.
At the '2026 International Risk and Insurance Workshop' hosted by the Korea Insurance Research Institute at the FKI Tower in Seoul, Arthur Charpentier, a professor at Quebec University, stated, "The goal of actuarial science is not to predict who will have an accident but to estimate the probability of accidents fairly. No matter how much data is collected, some uncertainty remains."
Charpentier noted that AI and big data could improve the accuracy of risk assessments, leading to more precise premium calculations. However, he cautioned that overly granular risk segmentation could weaken the insurance principle of risk diversification, potentially resulting in high-risk individuals facing higher premiums or even being unable to obtain insurance altogether.
Concerns about discrimination in AI-based premium calculations were also highlighted. Charpentier explained, "Simply removing sensitive information is not enough, as other data can still allow for inferences about protected characteristics such as gender or race."
Climate risk was cited as a prominent example. In France, enhanced natural disaster risk mapping has allowed for more detailed assessments of disaster risks in specific areas, leading some insurers to shy away from underwriting in high-risk regions. He emphasized, "Insurance is ultimately an industry based on trust," and stressed the importance of fairness and transparency in the use of AI.
Consumer perceptions of AI in insurance are mixed. Alex Jia, a professor at Peking University, shared findings from a survey conducted by the International Insurance Association, which included 6,000 insurance consumers across six countries. He noted, "While consumers view AI positively for insurance comparisons and claims support, they are negative about its use in personalized advertising." Jia added, "Consumers prioritize data security, information accuracy, and the ability to connect with human agents over the efficiency of AI."
AI itself may pose new risks that insurers need to manage. Charpentier assessed, "AI is not a completely new risk but a technology that amplifies existing risks." Jia also remarked that the AI insurance market is at a stage similar to the early cyber insurance market, with companies indicating a willingness to pay an additional 10% to 20% on current premiums for AI risk coverage.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.