Generative AI in Physical Education: A PLS-SEM Approach Among Preservice PETE Students in Taiwan

Abstract

Purpose : This study examined factors influencing preservice physical education teachers’ adoption of ChatGPT, extending the technology acceptance model with trust and perceived risk. Method : Participants were 345 physical education teacher education students in Taiwan who completed measures of perceived ease of use, perceived usefulness, trust, perceived risk, and usage behavior. Usage was operationalized across three authentic tasks, information gathering, problem solving, and idea generation to reflect ecological validity. Data were analyzed using partial least-squares structural equation modeling. Results : Perceived ease of use positively predicted perceived usefulness and usage, and perceived usefulness positively influenced both trust and usage. Perceived risk negatively predicted trust, and trust positively predicted usage. The model explained 78.4% of the variance in trust and 72.1% in usage behavior. Discussion/Conclusion : Findings highlight trust and risk as key affective–evaluative mechanisms in embodied learning, offering guidance for responsible integration of generative AI into physical education teacher education.