The impact of an AIPER on health behavior improvement among sedentary adults: a longitudinal extension of the TAM

Abstract

Background: Sedentary behavior has become a major public health concern and is closely associated with various unhealthy behaviors and chronic diseases. Artificial intelligence shows promise for promoting health behavior change through personalized exercise interventions. Objective: To examine, over a 6-month period, the effect of an AI-based Personalized Exercise Recommendation System (AIPER) on Health Behavior Improvement (HBI) among sedentary adults and to analyze its psychological mechanisms. Methods: A two-wave survey (T1 and T2, 6 months apart) was conducted with 492 sedentary participants. Measures covered TAM-related behavioral variables—Perceived Ease of Use (PEOU), Perceived Usefulness (PU), Attitude Toward Use (ATU), Behavioral Intention (BI), and System Use (SU)—as well as Health Self-Efficacy (HSE), Health Behavior Improvement (HBI), and demographics. Factor analyses, correlation analyses, and structural equation modeling were performed using SPSS 23.0 and Amos 23.0. Results: AIPER significantly promoted HBI among sedentary adults. Chain mediation effects were identified, whereby PEOU and PU influenced ATU and BI, and together with HSE indirectly affected SU, ultimately improving HBI. Conclusion: AIPER can increase SU and indirectly improve HBI in sedentary populations. It is recommended that government agencies, enterprises, and universities/research institutes implement AI health-management systems, formulate individualized and scientifically grounded behavior-change plans, enhance the health behaviors of sedentary groups, and advance personalized, evidence-based, and sustainable public health promotion.