A Study on Korean EFL Students’ Perceptions of ChatGPT Feedback in Argumentative Writing
2
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This study explores how Korean EFL university students perceive ChatGPT-generated feedback in the context of argumentative writing, focusing not only on its surface-level utility but also on its emotional and pedagogical impact. Using a mixed-methods design with a quantitative emphasis, 31 students completed two writing tasks, revised them using ChatGPT, and reflected on their experience through a structured survey and reflective journal. The findings reveal a clear preference for feedback related to grammar and vocabulary, while students were more skeptical of higher-order suggestions, such as content development or idea organization. Correlational analysis showed that students’ emotional responses were strongly tied to their perceptions of technical reliability and functional usefulness. While learners appreciated ChatGPT’s efficiency, they expressed concerns about its accuracy and long-term impact on writing autonomy. Thematic analysis of reflections underscored a dual perception of ChatGPT as both a helpful assistant and a source of ethical tension. These insights highlight the need to move beyond binary views of AI in education and instead focus on supporting learners in critically evaluating, adapting, and responsibly integrating AI-generated feedback into their writing processes.
Abstract
This study explores how Korean EFL university students perceive ChatGPT-generated feedback in the context of argumentative writing, focusing not only on its surface-level utility but also on its emotional and pedagogical impact. Using a mixed-methods design with a quantitative emphasis, 31 students completed two writing tasks, revised them using ChatGPT, and reflected on their experience through a structured survey and reflective journal. The findings reveal a clear preference for feedback related to grammar and vocabulary, while students were more skeptical of higher-order suggestions, such as content development or idea organization. Correlational analysis showed that students’ emotional responses were strongly tied to their perceptions of technical reliability and functional usefulness. While learners appreciated ChatGPT’s efficiency, they expressed concerns about its accuracy and long-term impact on writing autonomy. Thematic analysis of reflections underscored a dual perception of ChatGPT as both a helpful assistant and a source of ethical tension. These insights highlight the need to move beyond binary views of AI in education and instead focus on supporting learners in critically evaluating, adapting, and responsibly integrating AI-generated feedback into their writing processes.