Real-time AI-mediated peer feedback in L2 writing: a mixed-methods study of learner engagement and accuracy development

Although interest in integrating Artificial Intelligence (AI) into language learning has increased, research examining the distinct effects of AI-mediated peer feedback (AIMPF) and traditional peer feedback (TPF) on L2 writing improvement remains limited. To address this gap in the literature, the present study explored the effectiveness of AIMPF in comparison to TPF in enhancing the accuracy and engagement (e.g. vigor, dedication, and absorption) in the writing of Iranian MA English students. Sixty intermediate Iranian English students enrolled in an academic writing course at a university in Iran were selected using purposive sampling and then randomly assigned to either a TPF group or an AIMPF group. The AIMPF group utilized an AI-supported feedback tool alongside peer review, while the TPF group received TPF sessions. Both groups completed pre- and post-tests, which included writing tasks and an engagement questionnaire. In addition, participants in the AIMPF group completed an open-ended survey to gather their attitudes and perceptions regarding AI-facilitated feedback. Results indicated that both groups demonstrated improvement in writing accuracy and engagement; however, the AIMPF group exhibited significantly greater gains in both areas. Qualitative data revealed that participants in the AIMPF group found AI-supported feedback beneficial for identifying language errors and supporting their revision process, although they also expressed reservations concerning the absence of human interaction within the feedback experience. These findings support that AI-supported peer feedback holds promise as a valuable method for developing L2 writing ability, potentially fostering both increased engagement and improved accuracy.