A286: Deepseek Empowering Extracurricular Physical Activities

Traditional management approaches for school extracurricular physical activities increasingly struggle with efficiency and personalization amid growing program diversity. This study investigates DeepSeek—a domestically developed AI large language model—to optimize activity design, event logistics, and training strategies in physical education (PE) through human-AI collaboration. Utilizing a qualitative design, semi-structured interviews were conducted with 15 PE teachers across Beijing secondary schools. Case studies of DeepSeek-assisted programs (e.g., AI-curated recess exercise sequences, data-driven sports festival schedules) were analyzed via thematic coding. Triangulation of interview transcripts, teacher journals, and activity logs ensured methodological rigor. Three key themes emerged: 1) Personalized activity recommendations (e.g., blending dance and strength training for fitness gaps) increased student participation by 68% based on teacher estimates; 2) AI-generated scheduling reduced time spent on logistics by 40%, though 87% of teachers modified outputs for local constraints; 3) Performance analytics enabled tiered skill development plans, yet 93% emphasized manual oversight for high-risk activities. Teachers particularly valued AI's ability to synthesize student health data with curriculum standards but noted limitations in adapting to sudden environmental changes. DeepSeek effectively augments PE management through data-informed automation while requiring strategic human intervention for contextual adaptation. The study establishes a proof-of-concept for AI's role in reducing administrative burdens and enabling differentiated instruction, contingent on dynamic teacher-AI workflows. Future research should explore hybrid decision-making models and resource-sensitive implementation protocols for varied educational contexts.