Student Engagement with genAI Feedback in L2 Writing: Evidence from Chinese Students’ Engagement with Kimi Feedback

Despite substantial literature on the application of generative AI (genAI) feedback in L2 students’ writing process, little has explored how students engage with such humanoid feedback. Informed by the multi-dimensional conceptual framework of student engagement, this study explores Chinese EFL students’ affective, cognitive, and behavioral involvement in Kimi feedback. Multiple sources of data including archived Kimi records, multiple drafts, reflection papers, questionnaire, and interview were analyzed quantitatively and qualitatively. The results revealed the students exhibited overall positive engagement with Kimi’s multi-leveled feedback. Affectively, they exhibited interest in Kimi feedback, happy to receive it, willing to negotiate with Kimi, and acknowledged the value of Kimi feedback in improving writing quality. Cognitively, they understood most Kimi feedback, deployed various cognitive and metacognitive operations to process the feedback, decide, and evaluate its incorporation. Behavioral engagement was found in high adoption of accurate Kimi feedback as well as dismissal of most inaccurate feedback. Meanwhile, behavioral engagement sometimes might emerge without sufficient cognitive engagement. The research provides evidence for the multi-faceted nature of student engagement with genAI feedback in L2 writing activity. It also highlights a multi-faceted lens for teachers to scaffold students for more effective utilization of such affordance.