This study examines abstract-level research trends in AI-mediated English-speaking education through a systematic text-mining analysis of English-language abstracts drawn from 130 Korean journal articles and theses published between 2015 and 2025. The dataset was constructed by retrieving English-language abstracts from RISS, a major Korean academic database and was analyzed using a combination of word-frequency analysis, TF-IDF weighting, N-gram extraction, and co-occurrence network analysis. These quantitative procedures were supplemented by topic modeling and CONCOR clustering to identify dominant thematic structures across the corpus. The results indicate that chatbot-based technologies—most notably PengTalk and ChatGPT—have emerged as the primary platforms for AI-mediated speaking practice, particularly within classroom-oriented instructional designs emphasizing proficiency development, level-based assessment, and feedback provision. Four major thematic domains were identified: classroom-based outcome validation, AI/chatbot-mediated support and proficiency gains, learner-centered task design and interaction, and speaker or user experience. Across these domains, AI is increasingly positioned not merely as an auxiliary tool but as an interactive learning partner that reshapes pedagogical practices and assessment approaches. The findings further suggest a growing emphasis on engagement, interactional authenticity, and formative feedback in AI-mediated speaking research. This study concludes by discussing implications for future classroom-embedded longitudinal research and the development of multimodal (speech–text) evaluation frameworks that foreground fairness, authenticity, and personalization in English-speaking assessment.
Building similarity graph...
Analyzing shared references across papers
Loading...
Yena Lee
Hye-Rang Om
Korean Journal of English Language and Linguistics
Building similarity graph...
Analyzing shared references across papers
Loading...
Lee et al. (Fri,) studied this question.
www.synapsesocial.com/papers/69a75f8fc6e9836116a2b063 — DOI: https://doi.org/10.15738/kjell.26..202601.84