AI-enabled learning analytics use relates to physical literacy and engagement in university PE via smart teaching and...

Digital transformation and AI-enabled learning analytics are reshaping higher education, and wearable-enabled analytics are increasingly used in embodied curricula such as university physical education (PE), but empirical evidence linking these systems to physical literacy remains limited. This study investigates how AI-enabled learning analytics use (wearable-derived dashboards and automated alerts) in smart PE relate to students' physical literacy and learning engagement, and tests whether perceived smart teaching quality and personalised feedback mediate these associations within an established human-centred learning analytics perspective. An explanatory sequential mixed-methods design combined a survey of 1,182 students from four Chinese universities with semi-structured interviews with 12 students and six staff members. Structural equation modelling showed that analytics use was associated with perceived smart teaching quality (β = 0.47, p < .001) and personalised feedback (β = 0.39, p < .001), which were in turn related to physical literacy (β = 0.28 and β = 0.36, respectively, p < .001) and learning engagement (β = 0.24 and β = 0.31, respectively, p < .001); direct paths from analytics use to physical literacy (β = 0.06, p = .080) and engagement (β = 0.05, p = .110) were small and not statistically significant, while bias-corrected bootstrap mediation estimates (5,000 resamples) indicated that the association operated primarily through teaching and feedback processes. Thematic analysis showed that students and instructors experienced analytics both as a "mirror and coach" and as a source of pressure, fairness concerns and heightened bodily visibility, with system reliability, assessment regimes and data literacy shaping these interpretations. Overall, the findings suggest that AI-enabled analytics are more consistently linked with physical literacy and engagement through pedagogical and feedback processes rather than through data exposure alone. By applying and testing established human-centred learning analytics mechanisms in a compulsory university PE setting, the study provides mixed-methods evidence to inform the design of smart PE initiatives that support physical literacy in higher education. Because the survey data are cross-sectional and self-reported, common-method bias cannot be fully ruled out and findings should be interpreted as associations rather than causal effects.