2025 NAIC Health Insurance Artificial Intelligence/Machine Learning Survey Results

EXECUTIVE SUMMARY

The Health Insurance Artificial Intelligence/Machine Learning Survey Results report aims to provide a
comprehensive understanding of the use of AI/ML by health insurers, the role of third-party
components, AI governance frameworks, and the alignment with NAIC AI Principles. The survey,
conducted by 16 states, gathered responses from 93 companies, indicating that 84% of health insurers use AI/ML across various product lines, including Individual Major Medical, Group Major Medical, and Student Health Plans.

Companies selling individual major medical health insurance are currently using or exploring the use of AI/ML primarily for utilization management practices (71%), disease management programs (61%), prior authorization for approval processes (68%), claims fraud detection (50%), for medical provider fraud detection (51%), and sales and marketing solutions (45%) for enhancing online sales, quoting, or
shopping experiences. Only about 4% of health insurers are using AI/ML to detect smoking and even
fewer insurers use facial recognition or behavior models to detect fraud. 12% of companies use AI for
denying prior authorizations and 14% of companies use AI to infer sensitive data, such as race or other
data values. 55% of health insurers use third-party components in their AI/ML Systems, 15% rely entirely on third-party AI/ML solutions, 13% use a combination of internal and third-party data and/or AI/ML components, and 10% develop AI/ML solutions internally.

Many companies have adopted principles focusing on accountability, transparency, security, and
privacy. The survey shows that many companies employ various methods to test for drift, bias, and
unfair discrimination in AI to include cross validation for accuracy, exploratory data analysis (EDA), to
analyze data for completeness and consistency, tracking performance metrics such as AUC, F-score,
confusion matrix, conducting equity audits, compliance audits, performance audits, and human
intervention in AI-driven decisions. Overall, while health insurers are taking steps to govern AI usage,
further analysis of this survey may provide insight into the next steps for regulatory frameworks and
industry practices to ensure that AI/ML technologies are used responsibly and ethically.