ABSTRACT This study examines how teacher interpersonal behaviours and AI attributes (responsiveness and anthropomorphism) influence students' willingness to communicate (WTC) with AI through the parallel mediators of classroom enjoyment and AI trust. Using survey data from 557 Chinese undergraduates in AI‐enhanced language classrooms, structural equation modelling revealed that teacher–student interaction predicts communication willingness both directly and indirectly through dual mediators. While responsiveness operates fully through indirect pathways, anthropomorphism exerts both direct and indirect effects. Crucially, classroom enjoyment enhances AI trust, confirming a synergistic affective‐cognitive sequence. Findings demonstrate teachers' irreplaceable role in AI‐enabled classrooms, highlighting human–AI synergy rather than substitution. This dual‐mediation perspective offers valuable insights for integrating AI while preserving essential human elements in education. This study examines how teacher interpersonal behaviours and two key AI attributes—responsiveness (speed/fluency of AI feedback) and anthropomorphism (human‐like cues of AI)—shape students' willingness to communicate with AI (WTCA) through the parallel mediators of classroom enjoyment and AI trust. Survey data from 557 Chinese undergraduates in AI‐enhanced language classrooms were analysed with structural equation modelling. Teacher–student interaction predicted WTCA both directly and indirectly via enjoyment and trust. AI responsiveness operated only through these indirect paths, whereas anthropomorphism exerted additional direct effects. Enjoyment, in turn, enhanced AI trust, confirming an affective‐cognitive synergy. Findings underscore teachers' irreplaceable role in human–AI collaborative learning and offer practical guidance for integrating AI without sacrificing essential human elements.
Zhang et al. (Sat,) studied this question.