Mechanism and Model Construction of “AI + Teacher” Collaborative Teaching in College English Blended Classrooms

Artificial intelligence is reshaping the landscape of college English teaching. However, current blended teaching mostly remains at the superficial integration stage of “online resources plus offline instruction”, with a lack of genuine collaboration between AI and teachers. This paper aims to construct the internal mechanism and operational model of AI-teacher collaborative teaching, rather than merely exploring its application effects. Based on collaborative teaching theory, activity theory and distributed cognition theory, this paper constructs a four-dimensional analytical framework for AI-teacher collaborative teaching, and systematically reveals the internal structure of the collaboration mechanism from four perspectives: task allocation, information interaction, decision coordination and feedback integration. On this basis, this study designs a three-stage collaborative teaching process model, and defines the human-AI collaboration nodes and interaction logic in the pre-class, in-class and post-class phases. It further identifies four major dilemmas that must be addressed in the construction of collaborative teaching models: the trade-off between efficiency and in-depth learning, the conflict between standardization and personalization, the paradox between learner autonomy and over-reliance, and the tension between certainty and uncertainty. Corresponding solutions are also proposed accordingly. The research reveals that AI-teacher collaborative teaching is essentially not a simple superposition of technical functions, but a reconstruction of human-AI relationships. The establishment of effective collaboration mechanisms requires joint support from four dimensions: technology, teachers, students and institutional mechanisms. This paper provides a systematic framework for the theoretical deepening and model design of human-AI collaboration in college English blended teaching.