Profiles and transitions of psychological adaptation in AI-assisted Japanese language learning

As artificial intelligence becomes increasingly integrated into language learning, learners may differ in how they adapt psychologically to AI-assisted study. This study examined profiles and transitions of psychological adaptation to AI-assisted Japanese language learning and their associations with learning self-efficacy and burnout. A three-wave longitudinal survey was conducted across one academic semester, with 457 learners at Time 1. Technostress and psychological resilience served as indicators in latent profile and latent profile transition analyses. Three profiles were consistently identified: maladaptive adaptation, moderate adaptation, and positive adaptation. The maladaptive adaptation profile decreased from 26.48% at Time 1 to 17.74% at Time 3, whereas the positive adaptation profile increased from 27.35% to 35.73%. The positive adaptation profile showed the greatest temporal stability, while learners in the maladaptive adaptation profile most commonly transitioned to moderate adaptation. At Time 3, learners in the positive adaptation profile reported the highest Japanese learning self-efficacy and the lowest learning burnout, whereas those in the maladaptive adaptation profile showed the reverse pattern. By integrating the Job Demands-Resources framework with a person-centered longitudinal approach, this study demonstrates that psychological adaptation to AI-assisted language learning is heterogeneous, dynamic, and meaningfully associated with learners' confidence and burnout.