HOME

Q. You recently published two papers in SSCI-indexed journals and were recognized by the Office of Research as part of its “Outstanding Papers of the Month” selection. Congratulations! Could you introduce the key findings of these two papers and explain their significance?

A.  Paper 1: Examining the Thematic Structure and Temporal Reconfiguration of The Asia-Pacific Education Researcher (TAPER): A BERTopic Analysis

This study examines the research topics covered by The Asia-Pacific Education Researcher (TAPER), a leading education journal in the Asia-Pacific region, over the past 13 years and how these research trends have evolved over time.

To overcome the limitations of traditional word frequency-based analyses, I used the latest BERTopic technique to analyze the abstracts of 962 articles published in TAPER, identifying 13 major research topics.

The study yielded two key findings. First, research on emotions and motivation, teachers’ emotional labor, and feedback literacy has steadily increased over the past decade. Second, following the COVID-19 pandemic, the journal's research landscape underwent a "reconfiguration" rather than a simple shift. While research on technology acceptance and STEM declined, studies on emotional labor and feedback literacy expanded significantly.

This study goes beyond analyzing a single journal by examining how academic journals accumulate and reorganize knowledge. It offers valuable insights to help researchers identify future research directions and journal editors assess their research coverage and explore opportunities for further development.

Paper 2: ChatGPT for Automated Writing Evaluation: Scoring and Feedback Across Prompt Conditions

This study examines how reliably ChatGPT can function as an automated writing evaluation (AWE) tool for English writing under different prompt design strategies. I configured six ChatGPT models using combinations of five prompting strategies and compared their evaluations of 60 English essays written by Korean university students with those of human raters. Using the many-facet Rasch model, I analyzed scoring consistency, severity, and bias.

The analysis revealed that a prompt combining Chain-of-Thought (CoT) and Fill-in-the-blank strategies achieved scoring consistency comparable to that of human raters, while Persona and Few-shot strategies helped reduce excessive leniency in scoring. However, differences from human raters remained in higher-order assessments, such as argumentation, and learners showed the strongest preference for specific, balanced feedback.

The significance of this study lies in demonstrating that, rather than viewing ChatGPT as an all-purpose tool or dismissing it as unreliable, it can serve as an effective supplementary tool for supporting instructors' feedback in formative assessment when appropriate prompt design is combined with instructor oversight, even without additional modifications such as fine-tuning.

Q. Your research to date has focused on three main areas: ① AI-based English language education, ② analysis of English language learners’ psychological factors, learning strategies, and behaviors, and ③ educational assessment, measurement, and instrument development. With the recent emergence of generative AI rapidly transforming the educational landscape, how do you envision these three research areas converging and evolving in the future?

A.  While it is difficult to provide a definitive answer to this question, I believe these research areas will converge in three main directions.

First, research examining the interaction between AI and learners itself will become increasingly important. While previous studies have focused primarily on prompting strategies or individual factors, I expect more research to comprehensively analyze how the processes and patterns of interaction with AI influence English language learning outcomes. This line of research is closely connected not only to AI-based English language education but also to educational assessment and measurement. If outcomes vary depending on learners’ ability to use AI, AI-assisted assessments may end up measuring AI proficiency rather than English language proficiency.

Second, research on how learners perceive AI during the collaborative processes will also become increasingly important. Understanding how learners perceive AI's role and the authorship of AI-generated work, as well as how their beliefs and attitudes toward AI influence learning, will be essential for developing effective AI education and AI ethics education.

Finally, I believe these lines of research will ultimately lead to the development of reliable and valid measurement instruments. In fact, instruments for measuring learners’ acceptance of generative AI feedback are already being developed, and I am also planning to develop an instrument to measure learners’ beliefs about AI and AI-generated outputs. Looking ahead, I expect continued development of instruments to measure AI literacy and prompt literacy, along with growing research into how these competencies relate to English language learning outcomes.

Q.  I understand that you have been consistently conducting research on AI-based English language education since before AI became a key technology in education. What initially inspired you to pursue research in this field, and what has motivated you to continue your research over the years?

A.  My interest in research methodology was the main reason I began studying AI-based English language education. After earning my Ph.D. in English Language Education, I enrolled in a master's program in Big Data to deepen my understanding of quantitative research methods. During the program, I was introduced to machine learning algorithms and AI technologies, which led me to explore how they could be applied to English language education. Looking back, I believe those early explorations laid the foundation for my current research.

The big data analysis techniques and AI-related methodologies I learned at the time became an important foundation for my subsequent research. That experience has been particularly valuable for my study on a predictive model for English academic achievement, which is currently under review, as well as my recently published BERTopic-based analysis of TAPER.

What has kept me engaged in this research is my fascination with understanding how AI technologies work, along with the rapid pace of development in this field. Continuously learning about new AI technologies and exploring how to integrate them into English language education research remains my greatest motivation for pursuing this work.

Q.  Your research has covered a wide range of areas, including AI-based English language education, educational assessment, and the analysis of learners' psychological factors and behaviors. You also serve as Director of the Glocal Communication Competency Center and Head of English Education at the Hyangseol Institute for Liberal Arts Education Research. Looking ahead, are there any interdisciplinary research topics or collaborative research directions you would like to pursue with researchers from other disciplines at Soonchunhyang University, particularly in the areas of AI, education, language, and data analysis?

 A. Just thinking about the possibilities fills me with excitement. Looking ahead, I would like to pursue interdisciplinary research in three main directions.

First, I would like to collaborate with researchers in computer engineering and AI to expand beyond prompt optimization into fine-tuning and the development of lightweight AI models. Building on this work, I hope to develop AI models tailored to Soonchunhyang University's educational environment and analyze students' learning behaviors to create an AI literacy framework unique to the university.

Second, I would like to collaborate with researchers in psychology and education to investigate how interactions with AI affect learners' emotions, motivation, and self-efficacy. Given that my recent TAPER analysis also identified a growing trend in research on emotions and motivation, I believe this is a timely and academically significant research topic.

Finally, I would also like to actively participate in collaborative research within the university to standardize AI literacy assessment criteria and measurement instruments for use in general education and major-specific courses. I hope that the findings from these studies will provide a meaningful foundation for developing an AI-based educational model unique to Soonchunhyang University.

Q. Is there anything else you would like to share?

 A. It is a great honor to have this opportunity to share my research. Looking back, all my research has ultimately centered on one question: "How can we help our students learn English more effectively and develop the communication skills they need in today's world?"

Moving forward, I intend to continue using this question as the starting point for my research, creating a virtuous cycle in which challenges encountered in educational settings inform my research, and insights gained from that research are applied back to educational practice. Through these efforts, I will continue contributing to the development of Soonchunhyang University students' English proficiency and the communication skills they need in the age of AI. Thank you.