ChatGPT의 피드백을 통한 한국어 작문 교육－교사 피드백 비교를 통하여－

초록

This study aims to examine the characteristics of teacher and ChatGPT feedback in writing, investigate learners’ perceptions of feedback and their usage patterns of ChatGPT, and explore future applications based on these findings. The results indicate that teacher feedback primarily focuses on the linguistic aspects of writing, particularly linguistic accuracy, whereas ChatGPT provides a relatively more balanced approach by addressing language, content, and structure with greater depth and detail. In terms of strategy, teachers predominantly use direct correction strategies, resulting in relatively simple forms of feedback. In contrast, ChatGPT feedback follows a three-part macroscopic structure consisting of overall evaluation, revision suggestions with explanations, and a revised writing sample. At a more detailed level, ChatGPT employs strategies such as recasting, example references, and evaluative comments, enabling a more systematic and progressive approach. Survey results show that learners prefer clear and specific feedback, with 70% of respondents indicating a preference for ChatGPT feedback when comparing it with teacher feedback. Lastly, by incorporating the survey findings, this study proposes stage-specific applications of ChatGPT in writing education based on the strengths and weaknesses of both teacher and ChatGPT feedback, utilizing principles of prompt design.

키워드