Investigating the place and value of GenAI feedback in instructed ESL writing: A comparison with peer and teacher feedback

Presenter: Xiaosa Li (Nanjing University of Information Science & Technology, China)

This study examines the types, revision impact, and perceived value of feedback from three sources (GenAI, peers, and teacher) in an ESL writing context. Drawing on a recognition-based framework that distinguishes recognitive and extra-recognitive feedback, the study investigates: (1) what types of feedback students receive from each source, (2) what types of revisions they make in response, and (3) how they perceive the value, effectiveness, and motivational impact of different feedback sources. Data were collected from 36 junior English majors across multiple drafting rounds and analyzed through NVivo coding of feedback function and focus, revision operations and levels, and student survey responses. Results show that GenAI feedback was overwhelmingly form-focused and functionally dense, frequently bundling evaluation, explanation, and suggestion within single comments. Teacher feedback exhibited greater meaning-level engagement and a more distributed functional structure, while peer feedback was primarily affective-relational. Revision patterns shifted from surface-level substitutions in early drafts to increased meaning-level additions and deletions in later rounds. Students rated teacher feedback highest in value, effectiveness, and motivational impact, followed by peer feedback, with GenAI rated lowest. Findings highlight the complementary but distinct pedagogical roles of human and AI-generated feedback.

Xiaosa Li graduated from Nanjing University with a PhD degree in applied linguistics. She now teaches in Nanjing University of Information Science and Technology. Her research interests include L2 writing, feedback literacy, and translation didactics.