VietTESOL International Convention 2026

While recent studies on artificial intelligence in language education have largely focused on learners’ perceptions and attitudes, limited attention has been paid to how students actually respond to different sources of feedback during the revision process. Adopting a qualitative-dominant approach, this study aims to explore how EFL students revise their academic writing after receiving AI-generated feedback and teacher feedback in a tertiary context in Vietnam. 14 participants completed two argumentative essays and revised their essays based on either AI-generated feedback or teacher feedback. The data includes first essays, feedback comments, and revised essays. Drawing on revision analysis and written corrective feedback research, the study examined the types of revisions students make, the extent to which feedback was taken up, and the differences between revisions prompted by AI and those prompted by teachers. Results indicated that AI-generated feedback may facilitate more surface-level revisions related to grammar and language accuracy, whereas teacher feedback may encourage deeper meaning-level revisions involving idea development and organization. Additionally, while lower-proficiency students may prefer AI feedback, most students negotiated teacher feedback but accepted AI feedback automatically. This study is expected to contribute practical insights into the complementary roles of AI and teachers in supporting EFL writing development in higher education contexts.