Using actor-network theory to diagnose problematization failures in sports intelligent training systems: an AI-driven reconstruction approach

Sports intelligent training systems face a paradoxical promotion landscape: high expert acclaim but poor practical adoption. Rooted not in technical flaws but in flawed construction logic, this study adopts Actor-Network Theory (ANT), focusing on “problematization,” to reconceptualize system design. It identifies three core dilemmas: a consensus deficit in problem definition, insufficient evidence from real teaching contexts, and a lack of engaging translational narratives. These collectively undermine the system’s social legitimacy and impede its establishment as an Obligatory Passage Point (OPP). To overcome these barriers, an AI-augmented translation framework is proposed. This integrates three AI-driven mechanisms: NLP and topic modeling computationally synthesize a shared problem definition from multi-actor discourses; digital twins simulate real teaching ecologies to generate context-aware evidence, enhanced by Explainable AI (XAI) for transparency; and generative AI crafts personalized, multimodal narratives that align system benefits with the distinct interests of teachers, students, and administrators. This approach facilitates a fundamental shift from technology-driven to problem-driven construction logic. By enabling dynamic consensus building, robust evidence generation, and tailored value communication, the framework supports continuous alignment with the evolving actor-network. Thus, the system can be sustainably anchored as the essential pathway to resolving core contradictions in sports teaching.