Intelligent multimodal feedback and teacher-learner dynamics in GenAI-enhanced language classrooms

The rapid integration of generative artificial intelligence (GenAI) into language education has foregrounded the need to understand how intelligent feedback systems interact with teacher–learner dynamics to shape learning outcomes. This study examined the structural relationships among intelligent multimodal feedback quality, behavioral engagement, learner self-confidence and emotional state, trust and perceived fairness of AI feedback, teacher influence and classroom climate, teacher readiness for AI/GenAI integration, and L2 skill development in GenAI-enhanced EFL classrooms. Data were collected from 729 EFL learners using validated questionnaires and objective language performance measures. Following rigorous data screening, the measurement and structural models were tested using Structural Equation Modeling (SEM) with SPSS 27 and AMOS 24. The results indicated that intelligent multimodal feedback quality exerted significant direct and indirect effects on L2 skill development, primarily through learners’ behavioral engagement and emotional self-confidence. Trust and perceived fairness of AI feedback functioned as key mediating mechanisms, while teacher influence, classroom climate, and teacher readiness for AI integration significantly shaped the effectiveness of AI-supported feedback practices. The findings highlight that GenAI-enhanced language learning is best understood as a human-centered and relational process in which technological affordances, affective factors, and pedagogical conditions jointly contribute to learning outcomes. The study offers theoretical and empirical insights into the design, implementation, and evaluation of intelligent feedback systems in EFL contexts.