AI Health Services and Health Satisfaction Across Socioeconomic Groups in South Korea: National Cross-Sectional Study Using an Instrumental Variable Approach

Background: AI-enabled digital health services are rapidly expanding within health care systems and are expected to improve health management and access to health information. However, rigorous empirical evidence on whether AI health service use is associated with individual health satisfaction remains limited, particularly regarding whether these potential benefits differ across socioeconomic groups. Objective: This study aims to examine the relationship between AI health service use and health satisfaction and to assess whether this association varies across socioeconomic groups. Methods: Nationally representative data from the 2024 Digital Divide Survey in South Korea (N=15,000) were analyzed. To address potential endogeneity arising from self-selection and reverse causality, a 2-stage least squares instrumental variable approach was used. Robustness analyses using an alternative sample restriction, an alternative estimation method, and alternative instrumental variable specifications were conducted to assess the robustness of the findings. Subgroup analyses and interaction tests were conducted to assess socioeconomic heterogeneity. Results: <.001). The positive association was stronger among men, individuals living outside the capital area, those with lower income, people living alone, and individuals with disabilities. Conclusions: AI health service use was associated with higher levels of health satisfaction, and this association appeared stronger among several socioeconomically disadvantaged groups. These findings suggest that AI health services may function as complementary health resources and highlight the importance of considering socioeconomic heterogeneity when developing and evaluating AI-enabled health policies.