위햇님 Hatnim We

DOI:10.15702/mall.2025.28.2.1 Vol.28(No.2) 1-30, 2025

This study explored the potential of artificial intelligence (AI)-based tools in Korean reading fluency instruction for beginner learners whose first language is Japanese, and it analyzed their perceptions and attitudes toward such instruction. Microsoft’s ‘Reading Progress’ was implemented in an actual classroom setting, followed by a survey to examine the learners’ responses. The analysis revealed that learners experienced improvements in key components of reading fluency, such as pronunciation accuracy, sentence-level reading, and prosody. They also reported increased confidence and interest in reading activities. These findings suggest that AI-based reading activities can effectively support both the cognitive and affective aspects of learning. Based on these results, this study proposes three practical directions for designing AI-assisted Korean reading fluency instruction. First, the instructional design should align with the core components of fluency and enable the effective use of automatic feedback functions. Second, repetitive learning structures and flexible assessment systems that promote learner motivation should be incorporated. Third, it is important to ensure technical reliability and user accessibility through clear guidance and learner support. While this study is limited as it did not empirically verify instructional effectiveness and was restricted to a specific learner group, it contributes by presenting a concrete classroom-based case in an underexplored area of Korean language education. These findings are expected to serve as foundational data for future research and practice in AI-supported reading fluency instruction.

Korean language education, reading fluency, AI tools, Reading Progress, Japanese L1 learners, learner perception

Min Hye Cho

DOI:10.15702/mall.2025.28.2.31 Vol.28(No.2) 31-54, 2025

This study conducts a comparative analysis of English as a Foreign Language (EFL) textbooks used in the Korean-ethnic Joseonhakgyo (朝鮮学校) secondary schools and Japanese secondary schools. Chongryon (總聯), a pro-North Korean organization, has established and operated Joseonhakgyo with financial support from North Korea, allowing them to maintain independence from the Japanese government. Utilizing a text mining approach, textbooks published by Hagusobang (学友書房), a publishing company managed by Chongryon, were analyzed and compared to those of the Japanese publisher Gakko Tosho Kabushiki Gaisha (学校図書株式会社). This study aims to identify whether there is alignment between the EFL learning content of Joseonhakgyo and Japanese secondary schools, along with an examination of how Chongryon has responded to internal expectations―particularly from students and parents―who seek educational outcomes that facilitate integration into mainstream Japanese society. To achieve this, text mining techniques, like topic modeling and vocabulary analysis, were utilized to discover patterns in learning content and reveal how the textbooks differentiate. Findings reveal that the textbooks share similarities in the core of their educational content; however, there are clear differences regarding their educational focus. The Japanese textbook prioritizes a balanced development of the four language skills, whereas the Joseonhakgyo textbook places greater emphasis on ethnic identity through the inclusion of cultural elements. To better align with Japan’s mainstream curriculum, this study recommends Joseonhakgyo expand their use of real-life speaking and listening activities and diversify their reading and writing tasks by integrating more multicultural content alongside their ethnic cultural content. Additionally, the findings confirm that text mining is an effective tool for objectively analyzing educational materials, offering empirical support for curriculum development. Future studies should incorporate longitudinal analysis of curriculum changes and combine qualitative methods to gain deeper insights.

ethnic education, comparative analysis, EFL education, school textbooks, text-mining

Hyeyoung Jung

DOI:10.15702/mall.2025.28.2.55 Vol.28(No.2) 55-76, 2025

This study investigates the integration of AI-assisted student-generated materials (SGMs) in a test-driven high school English as a Foreign Language (EFL) listening classroom. It aims to reconcile institutional demands with the transformative potential of student-centered, technology-enhanced learning. Grounded in a sociomaterial framework, the research analyzes audio-recorded classroom interactions, post-survey responses, and teacher reflections to examine the impact of AI tools on student engagement and learning outcomes. Findings indicate increased student interest and motivation resulting from the use of AI-assisted SGMs. The study identifies four key elements for successful implementation: aligning SGMs with curricular goals, incorporating problem-based tasks to promote collaboration, evaluating multimodal content, and involving students in test creation to address linguistic features and institutional requirements. These strategies enhanced student participation and fostered a more interactive and reflective learning environment. By highlighting the balanced interplay among actors, tools, and contexts, where no single element dominates and all collectively contribute to the learning process, the study provides pedagogical insights for integrating AI tools into student-centered learning. It also sheds light on the evolving roles of teachers and supports the development of innovative practices in test-oriented language teaching contexts.

student-generated materials, artificial intelligence, sociomaterialism, listening, high school, EFL

Eunjin Lee

DOI:10.15702/mall.2025.28.2.77 Vol.28(No.2) 77-98, 2025

This study examines the content characteristics of and learner reactions to English language learning videos on YouTube by comparing content produced by native English-speaking and native Korean-speaking YouTubers. The study collected 56 videos from the top 20 English education channels and analyzed the content characteristics using a customized evaluation checklist. Additionally, 25,283 user comments were analyzed through text mining techniques. The results showed that both native English-speaking and native Korean-speaking YouTubers primarily employed drills and audio-lingual methods. However, native English speakers tended to use more cognitively engaging strategies such as reconstruction and situated approaches. Technological features such as subtitles and sound effects were commonly used; notably, native English speakers used visual images, while native Korean speakers provided video clips for practical pronunciation. Furthermore, text mining techniques were employed to examine viewers’ responses to the popular English-language content. The findings revealed that viewers appreciated the content and admired the YouTubers’ linguistic proficiency. In contrast, learners expressed different expectations and reactions to the content created by native English-speaking and native Korean-speaking YouTubers. Consequently, this study offers insights for developing more effective designs for English language content on YouTube.

English learning on YouTube, content evaluation, comment analysis, learner reactions, text mining