2026, Vol., No.35

The Journal of General Education 2025 KCI Impact Factor : 2.65

pISSN : 2465-7581
https://journal.kci.go.kr/davinci

The purpose of this study is to examine learners’ reactions to and perceptions of the use of generative AI in writing courses, and to explore both the acceptable scope of AI use and effective strategies for guiding and managing AI-assisted writing. To this end, surveys were administered three times in each of the first and second semesters of the 2025 school year to first-year Writing class learners. The results were then comparatively analyzed. A key finding from the first survey was that most learners firmly believed that humans should remain the primary agents of writing. In the second survey, the ‘AI Usage Guidelines’ proposed by the learners themselves indicated that they regarded ‘ethical compliance’ as the most important principle. In the third survey, conducted after AI use was partially permitted from weeks 9 to 14, most learners reported using AI and expressed the view that it is not merely a learning tool but also a powerful aid in the writing process. The results of the first through third surveys confirmed that, although the use of AI in writing may be permitted, appropriate guidance and supervision are necessary to ensure its proper use. Based on these findings, the proposed 'Scope of Use and Guidelines for Student Guidance and Supervision' are as follows. ① Increase the proportion of 'in-person writing' and 'oral presentations' ② Select topics that emphasize 'personal experiences' ③ Provide guidelines on the scope of AI use and its application ④ Require documentation of the 'writing process' ⑤ Encourage critical review and revision of AI-generated writing.

Recently, the importance of AI literacy has been emphasized, and discussions surrounding AI literacy education have been actively conducted. This thesis started with examining whether AI literacy education can be accepted in university writing classes and trying to find a specific plan. Referring to the scalability of the concept of literacy, there is an aspect that I leaned on the optimistic outlook for the intimacy between AI literacy and literacy. It was judged that AI literacy education would be sufficiently connected to literacy education, which has been proposed as an educational goal of writing subjects. After conducting the actual class, the educational effect related to AI literacy competency was reviewed. Overall, it is confirmed that the activities of preparing AI utilization reports and mutual evaluations had not a small educational effect. On the other hand, since the performance in terms of 'learning about AI' was low, it is difficult to evaluate that the goal of this class was that the link between 'learning with AI' and 'learning about AI' was sufficiently established. Another important point is that even if it can be evaluated that learners' AI literacy capabilities have been improved, it cannot be evaluated as an improvement in learners' literacy capabilities. The deepening of book blindness is a serious problem that must be taken as a serious problem in literacy education in writing classes. Therefore, the question of how AI literacy education will be accepted by writing education should begin with a more detailed review of the relationship between AI literacy and literacy.

The recent rapid proliferation of generative artificial intelligence (GAI) is bringing about significant changes not only to human creative activities but also to university writing education as a whole. Above all, the use of generative AI by learners in the process of university writing has become commonplace. Consequently, active discussions and research are underway in the field of university writing education regarding how to utilize this technology educationally and, in principle, prohibit its use. In this study, a survey and an analysis of writing assignments by text type were conducted on 63 students enrolled in the [Practical Writing] course at G University in Gyeongnam during the second semester of the 2025 academic year. Learners were instructed to write various practical texts directly during class time without the use of generative AI and submit them via the LMS, allowing them to utilize generative AI only during the revision stage. Through this process, this study explored learners' perceptions of generative AI and the educational potential of its use. Furthermore, it compared and analyzed the frequency and methods of generative AI utilization and changes in learners' perceptions during the revision process for each text type. The study results confirmed that learners generally perceive generative AI as a useful auxiliary tool during the revision stage. They showed a tendency to utilize generative AI at a relatively high rate during this phase. However, while the duration of AI usage varied by text type, it generally remained short-term, lasting less than 20 minutes, and differences in final satisfaction with the output were also observed. In particular, satisfaction with the effectiveness of AI was high during the revision of writing requiring structural clarity, whereas a relatively cautious attitude toward AI use was observed in writing where creativity or the expression of personal emotions is important. These results suggest that an educational model utilizing generative AI restrictively during the revision stage, premised on the learner's independent thinking, is more valid than indiscriminately applying it to the entire writing process. This is because it can serve as a method to prevent the "outsourcing of thinking."