Speaker
Description
The arrival of GenAI tools in language education has necessitated an understanding of their efficacy, particularly in propelling learners to greater writing proficiency. This study seeks to explore the extent to which Automated Writing Feedback (AWF) generated by ChatGPT can augment learners’ writing ability, together with their perceptions of its effectiveness, compared to the traditional teacher-led method. This 10-week, mixed-methods research investigated two groups: students who received feedback from the educator or from AI. Findings retrieved from quantitative data (students’ writing scores) indicate that both human and AI feedback can strengthen students’ writing performance, while qualitative data (survey results and interview responses) suggest that teacher feedback was superior in fostering students’ affective aspects. We conclude that AI holds the potential to assist the marking process by handling micro-level errors by supplying timely feedback, allowing educators to prioritize macro-level errors, and providing a more thorough feedback experience. Thus, this paper proposes that a blended feedback approach is adopted to maximize the students’ learning outcomes.
Keywords: Automated Writing Feedback, teacher feedback, writing performance, blended feedback approach
