Exploring Student Engagement in a Multiple-Interaction Environment: GAI, Peer, and Teacher Feedback in L2 Writing from an Ecological Perspective

Over the past decade, scholarly attention to how learners engage with feedback significantly grown within Second Language Acquisition (SLA) research. However, few studies have examined learner engagement with feedback in a multiple-interaction environment composed of generative artificial intelligence (GAI), peer, and teacher feedback. Grounded in the ecological affordance theory, the present study investigates the feedback engagement of eight non-English major undergraduates in a multiple-interaction environment within an EFL writing context. This study was carried out by analyzing written texts, questionnaires, stimulated recall interviews, and semi-structured interviews, with an emphasis on the behavioral, affective, and cognitive aspects of engagement. The findings reveal: (1) Three types of learner feedback engagement were identified: peer-teacher oriented type, GAI-teacher oriented type, and GAI-peer-teacher oriented type; (2) While correlations exist among the three dimensions of feedback engagement, discrepancies between cognitive engagement and affective or behavioral engagement were observed, particularly in GAI feedback stage; (3) GAI’s continuous mediation in some learners’ writing revision processes significantly influenced their engagement with the other two sources of feedback. These results offer both theoretical and practical implications for fostering students’ feedback literacy within intelligent education contexts and optimizing the design of multi-source feedback systems.