Engagement or disengagement: understanding the emotional dynamics of EFL learners in GenAI-assisted writing from the sociocultural perspective

Introduction Generative artificial intelligence (GenAI) has reshaped EFL writing education, yet the emotional dimensions of student engagement in GenAI-assisted writing remain underexplored. This longitudinal multiple-case study investigates how Chinese EFL learners' emotional engagement evolves during a GenAI-assisted college writing course and what factors drive this evolution. Methods Drawing on the sociocultural framework of L2 student engagement with written feedback, we traced the emotional trajectories of three focal students across three writing tasks (exposition, group feature article, and argumentation) within a 14-week course. Data were collected through semi-structured interviews with stimulated recall, reflective reports, GenAI-student chatlogs, drafts, informal chats, and course syllabus, and analyzed via thematic analysis. Results Our findings reveal three qualitatively distinct trajectories: positive-resilient, characterized by sustained positive affect with temporary fluctuations; pragmatic-transformative, marked by a gradual shift from practical disengagement to critical engagement; and vulnerable-contingent, moving from moderate disengagement, through tentative peer-mediated improvement, to eventual breakdown and intensified disengagement. These trajectories were shaped by the dialectical interplay between individual factors (e.g., English proficiency, GenAI literacy, mindsets, beliefs, and emotion regulation) and environmental factors (GenAI-relevant task design, peer and teacher mediation, and institutional policies). Discussion The study extends the sociocultural framework by identifying English proficiency and GenAI literacy as critical developmental competencies, embedding GenAI-related factors across multiple environmental levels, and foregrounding emotional engagement as a dynamic, fluid state that interacts with behavioral and cognitive engagement over time. The findings challenge universalist claims about GenAI's uniform affective benefits and underscore the need for differentiated pedagogical approaches that attend to learners' diverse emotional needs in GenAI-mediated writing contexts.