Abstract
This study investigates why artificial intelligence (AI) may hinder rather than enhance teaching efficiency in primary school physical education (PE). Guided by socio-technical systems theory, we conducted focus group interviews with 13 PE teachers (6 from Nanjing and 7 from Chongqing, China) who had at least three years of teaching experience and two years of AI implementation experience. Participants were purposefully selected through a two-stage sampling strategy: initial screening via open-ended questionnaires to identify teachers reporting negative experiences with AI integration, followed by snowball sampling to recruit additional participants with similar perspectives. Data collection employed a dual-facilitator approach using semi-structured interviews, with one moderator guiding discussions while another observed non-verbal cues. Qualitative content analysis revealed key barriers across four dimensions: technological (interface complexity, infrastructure limitations), employee (professional identity conflicts, interpersonal tensions), task-related (real-time monitoring challenges, reduced pedagogical flexibility), and organizational (inadequate support systems, unclear implementation policies). These findings suggest that successful AI integration in PE requires a holistic approach addressing both technological and human factors, rather than focusing solely on technological advancement. The study contributes to understanding how socio-technical interactions uniquely manifest in physically active learning environments.