Mediating generative artificial intelligence in sport pedagogy: lecturers’ pedagogical boundary-making in university physical education

The integration of generative artificial intelligence (GenAI) in higher education has been widely examined; however, less is known about how its pedagogical role is negotiated within practice-based disciplines such as physical education. This study examines how sport lecturers in Chinese universities mediate GenAI across teaching, feedback, and assessment practices. An interpretive qualitative design combined semi-structured interviews with fourteen lecturers and 28 classroom observations, with data analysed through reflexive thematic analysis. The findings show that GenAI was differentially positioned according to pedagogical purpose, task, and stage of learning. Lecturers positioned it selectively in preparation, conceptual support, planning, reflection, feedback, and, more tentatively, assessment-related judgement, while direct demonstration, observation and correction remained prominent in the performance-oriented practices examined. Tensions concerned learning depth, dependence, student autonomy, pedagogical regulation and assessment credibility, while equitable participation emerged as a further concern in some courses. The study contributes to sport pedagogy by conceptualising GenAI integration as a situated process of pedagogical boundary-making: the enacted conditions through which lecturers position GenAI within activity, specify what students must do with generated material, and reserve particular forms of learning for direct interaction.