Associations between physical education teachers' digital teaching competence and students' self-regulated physical activity in AI-augmented classes

This study examined associations between physical education (PE) teacher digital teaching competence and students' self-regulated physical activity in AI-augmented PE classes in China. Teacher digital teaching competence was defined as teachers' capacity to select and use digital resources for planning, instruction, assessment, feedback, learner support, and responsible professional practice; the measure did not assess artificial intelligence-specific technical knowledge or algorithmic expertise. An AI-augmented class was defined as a class using at least one system that automatically analyzed movement, activity, physiological, or performance data and generated individualized feedback, classifications, or recommendations used during instruction. Matched survey data were obtained from 140 PE teachers and 1640 students and aggregated to the teacher-class level after satisfactory within-class agreement and between-class reliability were established. Partial least squares structural equation modeling showed that teacher digital teaching competence was positively associated with students' self-regulated physical activity. Adaptive instruction in PE classes and perceived teacher learning support carried significant indirect associations. Perceived artificial intelligence support quality strengthened the first-stage paths from teacher digital teaching competence to both mediators. These findings identify general digital teaching competence, adaptive pedagogy, teacher support, and students' appraisal of artificial intelligence support as distinct components of the instructional ecology surrounding self-regulated physical activity. Because the data were cross-sectional, all paths are interpreted as associations rather than causal effects.