The impact of generative AI health knowledge acquisition on adolescents’ physical activity: the mediating role of exercise self-efficacy and the moderating effect of health information literacy

Background This study investigates the impact of generative artificial intelligence health knowledge acquisition (GAI-HKA) on adolescents’ physical activity (PA), with a focus on the mediating role of exercise self-efficacy (ESE) and the moderating role of health information literacy (HIL). Methods Using stratified cluster random sampling, a total of 2,306 valid questionnaires were collected from 10 secondary schools across five cities in southern China. Data were obtained using the self-developed Generative AI Health Knowledge Acquisition Scale, the Adolescent Exercise Self-Efficacy Scale, the Health Information Literacy Scale, and Liang’s revised Physical Activity Rating Scale. Structural equation modelling with PROCESS v3.5 macro (5,000 bootstraps) was employed to test direct, mediating, and moderating effects. Results GAI-HKA was significantly and positively associated with adolescents’ PA ( β = 0.376, p < 0.001). ESE partially mediated this relationship (indirect effect = 0.065, 95% CI [0.052, 0.079]). HIL positively moderated the pathway from GAI-HKA to ESE (interaction term β = 0.079, p < 0.001); the effect was more substantial among adolescents with high HIL ( β = 0.367) compared with those with low HIL ( β = 0.209). Conclusion GAI-HKA directly enhances adolescents’ PA and indirectly promotes exercise behavior through strengthening ESE. Moreover, higher HIL amplifies the positive effect of AI-based health knowledge on ESE. These findings suggest that schools and families should integrate efforts to cultivate adolescents’ information literacy and efficacy beliefs when promoting AI-based health interventions, thereby maximizing the benefits of digital technologies in fostering youth physical activity.