Content area
Abstract
ChatGPT, the newest pre-trained large language model, has recently attracted unprecedented worldwide attention. Its exceptional performance in understanding human language and completing a variety of tasks in a conversational way has led to heated discussions about its implications for and use in education. This exploratory study represents one of the first attempts to examine the possible role of ChatGPT in facilitating the teaching and learning of writing English as a Foreign Language (EFL). We examined ChatGPT’s potential to support EFL teachers’ feedback on students’ writing. To reach this goal, we first investigated ChatGPT’s performance in generating feedback on EFL students’ argumentative writing. Fifty English argumentative essays composed by Chinese undergraduate students were collected and used as feedback targets. ChatGPT and five Chinese EFL teachers offered feedback on the content, organisation, and language aspects of the essays. We compared ChatGPT- and teacher-generated feedback in terms of their amount and type. The results showed that ChatGPT produced a significantly larger amount of feedback than teachers and that compared with teacher feedback, which mainly focused on content-related and language-related issues, ChatGPT distributed its attention relatively equally among the three feedback foci (i.e., content, organisation, and language). Our results also indicated that ChatGPT and teachers displayed tendencies towards using different feedback types when evaluating different aspects of students’ writing. Additionally, we examined EFL teachers’ perceptions of using ChatGPT-generated feedback to support their own feedback. The five teachers reported both positive and negative perceptions of the features of ChatGPT feedback and the relation between ChatGPT and teacher feedback. To foster EFL students’ writing skills, we suggest that teachers collaborate with ChatGPT in generating feedback on student writing.






