Speakers
Description
The study examined how 62 Japanese university students engaged in an English summary‑writing activity that incorporated ChatGPT. Students drafted, received AI feedback, and revised their summaries. A separate post‑activity questionnaire captured their perceptions. Most students found ChatGPT helpful for identifying overlooked grammatical errors and improving organization and expression, though some were concerned about shifts in meaning. Overall, they selectively adopted AI feedback while maintaining their intended message.
Abstract section 4: Outcomes/results
In the questionnaire, about 90% of students reported that using ChatGPT in this activity was helpful, mainly because it accurately identified basic grammatical errors they often overlook. Students also valued suggestions for improving organization and flow, as well as more natural vocabulary and paraphrasing, which made writing easier and enhanced awareness of coherence and expression. However, some were concerned that AI revisions could alter meaning or replace words with ones they would not normally use, making the text feel less their own. Overall, students selectively adopted AI suggestions while prioritizing their original ideas and intended message.
Abstract section 3: Content/method
Sixty-two Japanese university students participated in a summary-writing activity incorporating ChatGPT. The activity involved three stages: drafting an initial summary, receiving feedback from ChatGPT, and revising the draft. Data were collected through a post-activity questionnaire. Quantitative data were analyzed by aggregating response distributions, while qualitative data were examined using thematic analysis to explore students’ perceptions and learning experiences.
Abstract section 2: Contribution/research questions
Research questions are:
1) How do students perceive the usefulness of feedback provided by ChatGPT?
2) What types of ChatGPT feedback do students accept?
3) What kinds of learning do they perceive they have gained through engagement with ChatGPT feedback?
Abstract section 5: References
Matsui, I. (2023). AI tsūru o katsuyō shite Nihonjin chūkōsei no gakushū no shutaisei o takameru koto o neratta raitingu shidō [Writing instruction aimed at enhancing Japanese secondary students’ learner autonomy through the use of AI tools]. EIKEN Bulletin, 35.
Yanase, Y. (2025). AI katsuyō ni yori eigo gakushūsha o jiritsuteki yūzā ni sodateru: Kyoto Daigaku no gakujutsu eigo raitingu jugyō ni tsuite no shōsatsuteki hōkoku [Fostering autonomous English users through AI use: A reflective report on an academic English writing course at Kyoto University]. Kyoto Daigaku Kokusai Kōtō Kyōikuin Kiyō, 8, 1–57.
Abstract section 1: Relevance
Although generative AI tools are increasingly used in education, little is known about how learners interpret and use AI-generated feedback. Research in Japan (Matsui, 2023; Yanase, 2025) shows that students can improve their writing with AI support; however, the ways they select, adapt, or reject AI suggestions—and how these decisions influence their learning and writerly agency—remain underexplored. This study examines how students perceive and engage with ChatGPT’s feedback and what they learn through this interaction. It aims to deepen our understanding of learner engagement with AI feedback and the development of student agency in writing.
| Title | How Students Evaluate and Use ChatGPT Feedback in Writing |
|---|---|
| Teaching Context | College and university education |