Emotion-Aware AI in Physical Education: Investigating Affective Computing's Role in Motivation, Regulation, and Self-Efficacy

Abstract

Understanding the role of emotion in student learning has become increasingly important in educational research, particularly in physically demanding disciplines such as Physical Education (PE), where motivation, confidence, and emotional resilience are critical for performance and engagement. Despite this, limited attention has been given to how artificial intelligence (AI), especially affective computing, can support these emotional and motivational processes within PE contexts. This study investigated the predictive power of AI-enabled affective computing on four key psychological constructs among undergraduate PE students: motivation, emotional regulation, academic self-efficacy, and control-value appraisals. A total of 409 PE students from Henan Province, China, participated. Structural equation modelling (SEM) was employed to examine the relationships between AI-driven emotional responsiveness and students' psychological outcomes in AI-supported PE learning environments. AI-enabled affective computing significantly and positively predicted all four variables. The strongest effect was observed for academic self-efficacy, followed by motivation, emotional regulation, and control-value appraisals. The SEM explained 67% of the variance in emotional regulation and 62% in self-efficacy, with robust model fit indices supporting the validity of the findings. These results highlight the potential of integrating emotionally responsive AI tools into PE programmes to create learner-centred, adaptive environments. Such integration can enhance emotional wellbeing, strengthen academic confidence, and promote sustained physical engagement among students.