Speaker
Description
The growing integration of generative AI in language education has raised important questions about how automated feedback can complement, rather than replace, human interaction in academic writing development. While both peer feedback and AI-generated feedback have been widely discussed, their combined pedagogical potential remains underexplored in EFL contexts.
This study investigates how different feedback sources contribute to the revision processes of postgraduate EFL learners in academic writing. Drawing on a mixed-methods approach, the research examines learners’ engagement with feedback, changes in writing quality, and their perceptions of AI-supported revision. Participants were involved in structured writing tasks and revision activities supported by either peer interaction or AI-generated feedback. Writing performance was evaluated using an analytic rubric, and quantitative results were complemented by learner reflections.
Findings indicate that AI-generated feedback supports surface-level improvements, particularly in grammatical accuracy and lexical choice, by providing immediate and consistent suggestions. In contrast, peer interaction fosters deeper engagement with meaning, including coherence, argument development, and audience awareness. Rather than functioning as competing approaches, the two feedback sources appear to play complementary roles in the writing process.
These findings highlight the importance of designing pedagogically balanced feedback practices that integrate both human and AI resources. The study contributes to current discussions on human–AI collaboration by offering insights into how feedback can be reconceptualized to support more effective and reflective academic writing development in EFL higher education contexts.
