A287: Innovation of School Physical Education Systems Driven by AI Technologies

Amidst educational digital transformation, school physical education faces challenges, including rigid teaching methods, vague evaluation standards, and insufficient personalized guidance. As a core next-generation technology, AI has transitioned from theoretical exploration to practical implementation in education. This study investigates the intersection of "AI + physical education" to establish a systematic theoretical framework elucidating the mechanisms and implementation pathways of AI technologies in school sports. By analyzing technology-enhanced education theories, it clarifies how AI addresses critical challenges, providing policymakers with theoretical foundations and offering schools methodological guidance for teaching reform. Combining theoretical construction and case validation, this study: 1) Conducted a systematic literature review of 136 AI-in-education publications to extract key technological features; 2) Analyzed 24 intelligent PE pilot projects across 12 Chinese provinces using NVivo12 coding; 3) Developed a "technology-education-PE" tripartite model via grounded theory (open/axial/selective coding), drawing on policy documents, technical white papers, and school evaluation reports. AI applications demonstrated three key characteristics: 1) Computer vision achieved 89.7% accuracy in skill assessment (vs. 67.3% for traditional methods); 2) Big data identified 78.5% of individualized motor skill development patterns; 3) Wearables increased activity monitoring coverage from 42% to 93%. The theoretical model revealed that AI transforms teaching through data-driven decisions, real-time feedback, and personalized adaptation. Successful projects shared three features: multimodal data fusion (87.5%), teacher co-design (79.2%), and iterative optimization (66.7%). This study advances smart education theory by: 1) Proposing a novel "dual-wheel" (data+pedagogy) model; 2) Defining stage-specific AI adaptation principles; 3) Creating an actionable implementation roadmap. While confirming Johnson's (2021) discipline-specific AI adaptation theory, it uniquely highlights PE's distinctive requirements. Practical contributions include a three-phase scaling strategy (pilot-refinement-expansion) and an AI solution evaluation framework (technical-educational-practical dimensions). Limitations involve sample size constraints, warranting expanded longitudinal studies. Future research should address ethical boundaries and AI's impact on teacher-student dynamics.