Understanding learner engagement in AI-mediated language learning: Growth mindset, resilience, self-efficacy, and AI-use behavior profiles

This study examined the relationships among growth mindset, resilience, self-efficacy, and learner engagement in AI-mediated language learning, while exploring variation in learners’ reported AI-use behaviors and experiences of AI-generated feedback. Drawing on survey data from 625 Malaysian undergraduates and follow-up interviews with 24 participants, the study employed structural equation modeling, latent profile analysis, and reflexive thematic analysis. The findings indicated positive associations of growth mindset and resilience with self-efficacy and learner engagement. Self-efficacy was positively associated with learner engagement and statistically mediated the associations of growth mindset and resilience with engagement. Latent profile analysis identified three patterns of reported AI-use behavior: Low AI-Use Behavior, Moderate AI-Use Behavior, and High AI-Use Behavior, based on AI-use frequency, revision cycles, revision depth, and feedback types used. Multigroup analyses indicated broadly comparable motivational relationships among Chinese and English language learners, although the growth mindset–engagement association was stronger among Chinese learners, while the resilience–self-efficacy association was stronger among English learners. The qualitative findings illustrated how participants representing the three profiles described interpreting, evaluating, and incorporating AI-generated feedback during language-learning activities. Participants’ accounts also reflected variation in how they assessed feedback, applied it in subsequent learning activities, and perceived the demands of AI-supported learning. These accounts provide contextual insight into learners’ experiences of AI-generated feedback without suggesting that profile membership determines such experiences. Overall, the findings indicate that self-efficacy was positively associated with engagement in the AI-mediated learning context examined, alongside heterogeneity in learners’ reported AI-use behaviors and experiences with AI-generated feedback.