Tracking foreign language speaking enjoyment across AI-mediated conversational topics: an idiodynamic perspective

Research on AI-integrated language learning has expanded rapidly due to the affordances AI offers in language classrooms. However, the role of AI in shaping learners’ emotions – particularly in skill-specific contexts and at a micro-temporal level – remains underexplored. This study examines how enjoyment dynamically fluctuates within and across AI-mediated conversational activities. Grounded in Positive Psychology (PP) and framed by Complex Dynamic Systems Theory (CDST), the study adopts a qualitative, case-based idiodynamic design. Two freshman pre-service English language teachers engaged in video-recorded, voice-based AI conversational activities. Data were collected through reflective journals, stimulated recall interviews, and video-mediated idiodynamic self-ratings to capture moment-to-moment and across-activity changes in enjoyment. The analysis focused on explaining enjoyment trajectories by linking temporal fluctuations to learners’ retrospective reflections. Findings showed that enjoyment was highly dynamic and context-sensitive, shaped by topic familiarity, expressiveness, activity demands, learner confidence, and AI interaction. More stable and elevated enjoyment emerged during familiar or self-selected activities, while pronounced peaks occurred when learners expressed opinions, shared personal stories, or engaged creatively. In contrast, enjoyment declined during cognitively demanding moments or when AI interventions disrupted interactional flow. Overall, the findings underscore the value of AI-mediated speaking activities in fostering positive emotional experiences, while highlighting the importance of topic selection and interactional design in shaping learners’ moment-to-moment enjoyment.