Purpose: The purpose of this study was (1) to examine the major thematic structures underlying the integration of artificial intelligence (AI) in physical education (PE) and (2) to identify the educational implications and future directions suggested by existing research. Method: Data were collected from studies published between 2000 and 2024 using the Web of Science, PubMed, EBSCO, and ERIC databases. The search combined technology-related terms?including AI, VR, AR, MR, and wearable technologies?with keywords related to physical education and PE instruction. Only studies that directly applied AI-related technologies to PE instruction or curricula were included. After screening procedures, a total of 282 English-language abstracts were selected for analysis. Following text preprocessing, a document?term matrix was constructed. Latent Dirichlet allocation (LDA) topic modeling and co-occurrence analysis were then conducted to identify the thematic structure and conceptual relationships within the research corpus. Results: The findings indicated that AI?PE research was organized around two major thematic streams.The first centered on “technological development,” encompassing algorithm design, model construction, motion recognition, and performance optimization. The second centered on “educational practice,” characterized by VR/AR-based learning environments, real-time feedback systems, intelligent assessment tools, and physical activity?related learning applications. Although these two streams remained relatively distinct, they were loosely connected through shared concepts such as “assessment” and “performance,” indicating potential points of convergence where technological advancements could be translated into instructional practice. Conclusion: Through text-based thematic network analysis, this study highlighted the need for future research to pursue a more balanced integration of technological, pedagogical, and institutional dimensions.
Liu et al. (Sat,) studied this question.