The Future of Education

Providing university students with feedback is widely accepted as an effective way to facilitate the development of L2 writing skills. Until recently, this process was largely reliant on feedback from instructors and peers. However, AI tools (e.g., ChatGPT) now offer a third option, potentially increasing the amount of feedback available to learners and decreasing the workload for instructors. Recent research has shown that while instructors still provide better quality feedback, AI feedback can increase learners’ motivation, self-efficacy, and engagement (Teng, 2024) as well as help them become more expressive writers. ChatGPT-generated feedback also induces higher affective, behavioral, and cognitive engagement among learners, compared to peer feedback (Huang & Teng, 2025). To investigate the ever-improving capabilities of ChatGPT, a qualitative study was conducted, framed by the research question: What are the differences between AI and instructor feedback on university students’ L2 essays? Twenty-four participants received feedback on a series of essays from the instructor and ChatGPT, which were analyzed utilizing Guo and Wang’s (2023) three foci as a framework for analysis: content, organization, and language aspects. Findings reveal that ChatGPT produced more detailed feedback than the instructor, with specific examples of how to improve existing sentence structures or how to develop existing references and content-related points. The instructor provided more feedback on elements that were missing from the essay drafts, as well as more overall critical and supportive forms of feedback. The implications of these findings and more will be discussed in this presentation with a special focus on L2 English writing instruction in higher education.