Jang Ho Lee
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jang Ho Lee has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
- Hansol Lee
Korea Military Academy
30 shared publications - Dongkwang Shin
Gwangju National University of Education
14 shared publications Dongkwang Shin
4 shared publications- Kyungmin Kim
Chung-Ang University
4 shared publications - Ernesto Macaro
Department of Education University of Oxford Oxford UK
3 shared publications - Heyoung Kim
University University
3 shared publications
Latest Publications
- Open Access
- 8 October 2026
- Published by Walter de Gruyter GmbH in International Review of Applied Linguistics in Language Teaching
Effects of speech-enabled LLM-based chatbots on L2 learning outcomes and affective domains: a multilevel meta-analysis
Recent developments in natural language processing and automatic speech recognition have significantly advanced chatbot communication using the voice mode, establishing these tools as viable resources for L2 speaking practice. More recently, the L2 chatbot domain has leveraged large language models (LLMs) to enhance contextual understanding and facilitate meaningful dialogues, prompting research into the effects of speech- and LLM-based chatbots on L2 learning outcomes and affective domains. This study presents a synthesis of these findings using a multilevel meta-analysis. Based on a dataset of 19 studies (32 samples, N = 2,643), the results revealed that the overall effect size for L2 learning outcomes was 0.68 ( p < 0.001). The overall effect size for affective domains was 0.64 ( p < 0.001). Moreover, moderator analyses indicated that a range of learner and intervention characteristics, including learner proficiency, educational level, interaction format, and context of chatbot use, significantly influenced the overall effectiveness.
- 8 October 2026
- Published by Walter de Gruyter GmbH in International Review of Applied Linguistics in Language Teaching
Open AccessEffects of speech-enabled LLM-based chatbots on L2 learning outcomes and affective domains: a multilevel meta-analysis
Recent developments in natural language processing and automatic speech recognition have significantly advanced chatbot communication using the voice mode, establishing these tools as viable resources for L2 speaking practice. More recently, the L2 chatbot domain has leveraged large language models (LLMs) to enhance contextual understanding and facilitate meaningful dialogues, prompting research into the effects of speech- and LLM-based chatbots on L2 learning outcomes and affective domains. This study presents a synthesis of these findings using a multilevel meta-analysis. Based on a dataset of 19 studies (32 samples, N = 2,643), the results revealed that the overall effect size for L2 learning outcomes was 0.68 ( p < 0.001). The overall effect size for affective domains was 0.64 ( p < 0.001). Moreover, moderator analyses indicated that a range of learner and intervention characteristics, including learner proficiency, educational level, interaction format, and context of chatbot use, significantly influenced the overall effectiveness.
- 4 September 2026
- Published by Wiley in TESOL Journal
- Vol. 17 (4)
Multimodal GenAI for Communicative Language Teaching: Gemini Live
This review explores the pedagogical affordances of Gemini Live, Google's multimodal generative AI (GenAI) chatbot, for communicative language teaching, with particular attention to task‐based interaction. Unlike earlier text‐ or voice‐based GenAI chatbots, Gemini 2.5 released in 2025 integrated live vision and voice to support context‐rich, goal‐oriented communication. It enables learners to engage in interactive, multimodal activities such as object‐based exchanges, gesture‐supported interaction, and screen‐shared collaboration. Although such activities do not automatically constitute full task‐based language teaching (TBLT), Gemini's multimodal environment offers a promising platform for scaffolded, authentic interaction, reflective post‐task discussion, and task‐supported language learning.
- 4 September 2026
- Published by Wiley in TESOL Journal
- Vol. 17 (4)
Multimodal GenAI for Communicative Language Teaching: Gemini Live
This review explores the pedagogical affordances of Gemini Live, Google's multimodal generative AI (GenAI) chatbot, for communicative language teaching, with particular attention to task‐based interaction. Unlike earlier text‐ or voice‐based GenAI chatbots, Gemini 2.5 released in 2025 integrated live vision and voice to support context‐rich, goal‐oriented communication. It enables learners to engage in interactive, multimodal activities such as object‐based exchanges, gesture‐supported interaction, and screen‐shared collaboration. Although such activities do not automatically constitute full task‐based language teaching (TBLT), Gemini's multimodal environment offers a promising platform for scaffolded, authentic interaction, reflective post‐task discussion, and task‐supported language learning.
- Open Access
- 18 June 2026
- Published by Cambridge University Press (CUP) in ReCALL
- Vol. 38 (3) , 1-16
Learners’ priority criteria for evaluating data-driven learning tools: An analytic hierarchy process study with traditional and GenAI-based concordancers
This study investigates how language learners prioritize criteria for evaluating data-driven learning (DDL) tools and how these priorities relate to tool preference. While previous research has proposed multiple criteria for evaluating the perceived effectiveness of DDL tools in capturing the emic dimension of the learner experience, empirical work has scantily examined how learners differentially prioritize these criteria, or how such priority structures can be empirically incorporated into systematic comparisons of tools. Given the recent emergence of generative artificial intelligence (GenAI) in language learning, this study examines how GenAI-based tools may contribute to DDL when judged against learner-valued effectiveness criteria. Drawing on a literature review, six criteria were identified: sentence comprehensibility, relevance to semantic learning, relevance to syntactic learning, perceived pedagogical value, accessibility for independent use, and support for autonomous learning. Thirty-five Korean EFL university students completed an analytic hierarchy process (AHP) task. First, they made pairwise comparisons among the six criteria, yielding priority weights that indicated the relative importance learners attach to each aspect of DDL effectiveness. Second, the study compared the perceived effectiveness of a traditional online concordancer that accesses the British National Corpus with that of a custom-developed GenAI-based concordancer, prioritizing these weighted criteria. The results indicate that learners assign differentiated importance to the six criteria and that these priority patterns are associated with the concordance they regard as more effective. This study demonstrates the usefulness of AHP for modelling multi-criteria, perception-based evaluations of DDL tools and incorporating learners’ priority judgements into the comparison of alternative tool designs.
- 18 June 2026
- Published by Cambridge University Press (CUP) in ReCALL
- Vol. 38 (3) , 1-16
Open AccessLearners’ priority criteria for evaluating data-driven learning tools: An analytic hierarchy process study with traditional and GenAI-based concordancers
This study investigates how language learners prioritize criteria for evaluating data-driven learning (DDL) tools and how these priorities relate to tool preference. While previous research has proposed multiple criteria for evaluating the perceived effectiveness of DDL tools in capturing the emic dimension of the learner experience, empirical work has scantily examined how learners differentially prioritize these criteria, or how such priority structures can be empirically incorporated into systematic comparisons of tools. Given the recent emergence of generative artificial intelligence (GenAI) in language learning, this study examines how GenAI-based tools may contribute to DDL when judged against learner-valued effectiveness criteria. Drawing on a literature review, six criteria were identified: sentence comprehensibility, relevance to semantic learning, relevance to syntactic learning, perceived pedagogical value, accessibility for independent use, and support for autonomous learning. Thirty-five Korean EFL university students completed an analytic hierarchy process (AHP) task. First, they made pairwise comparisons among the six criteria, yielding priority weights that indicated the relative importance learners attach to each aspect of DDL effectiveness. Second, the study compared the perceived effectiveness of a traditional online concordancer that accesses the British National Corpus with that of a custom-developed GenAI-based concordancer, prioritizing these weighted criteria. The results indicate that learners assign differentiated importance to the six criteria and that these priority patterns are associated with the concordance they regard as more effective. This study demonstrates the usefulness of AHP for modelling multi-criteria, perception-based evaluations of DDL tools and incorporating learners’ priority judgements into the comparison of alternative tool designs.
- Open Access
- 1 January 2026
- Published by National Foreign Language Resource Center in Language Learning & Technology
- Vol. 30 (1) , 1-19
Investigating the impact of chatbot design features on affective variables: A mixed-methods study
This study investigated how chatbots, designed with distinct personality traits (more engaging vs. less engaging) and feedback features (basic vs. enhanced) influenced L2 learners’ affective variables, such as anxiety, enjoyment, curiosity, and boredom, during their conversational interactions. Although second language acquisition research, including AI-based chatbot studies, has examined how learners’ emotions are shaped during L2 tasks, there has been insufficient focus on how variations in chatbot characteristics affect learners’ emotional experiences. To address this gap, we conducted a mixed-methods study of undergraduate EFL students, examining their interactions with three conversational agents, each featuring distinct combinations of personality traits and feedback features. Participants engaged in conventional tasks and completed a questionnaire to assess affective variables, which was supplemented with open-ended questions to gain further insights. Our findings revealed how L2 learners’ emotional experiences differ depending on the conversational agents’ characteristics. Based on these insights, we propose the pedagogical implications for the design and implementation of chatbots for L2 learning.
- 1 January 2026
- Published by National Foreign Language Resource Center in Language Learning & Technology
- Vol. 30 (1) , 1-19
Open AccessInvestigating the impact of chatbot design features on affective variables: A mixed-methods study
This study investigated how chatbots, designed with distinct personality traits (more engaging vs. less engaging) and feedback features (basic vs. enhanced) influenced L2 learners’ affective variables, such as anxiety, enjoyment, curiosity, and boredom, during their conversational interactions. Although second language acquisition research, including AI-based chatbot studies, has examined how learners’ emotions are shaped during L2 tasks, there has been insufficient focus on how variations in chatbot characteristics affect learners’ emotional experiences. To address this gap, we conducted a mixed-methods study of undergraduate EFL students, examining their interactions with three conversational agents, each featuring distinct combinations of personality traits and feedback features. Participants engaged in conventional tasks and completed a questionnaire to assess affective variables, which was supplemented with open-ended questions to gain further insights. Our findings revealed how L2 learners’ emotional experiences differ depending on the conversational agents’ characteristics. Based on these insights, we propose the pedagogical implications for the design and implementation of chatbots for L2 learning.
- 1 January 2026
- Published by National Foreign Language Resource Center in Language Learning & Technology
- Vol. 30 (1) , 1-21
GenAI-mediated activity theory (GMAT) for L2 teachers
This study proposes the GenAI-Mediated Activity Theory (GMAT) as a conceptual framework for understanding how generative AI (GenAI) reshapes second language (L2) teachers’ teaching preparation and practices within complex pedagogical ecosystems. Methodologically, the paper adopts a model-building approach. It recontextualizes and synthesizes conceptual elements from Engeström’s (1999) expanded Activity Theory (AT) to construct the model. Thus, the study illuminates the evolving role of L2 teachers as adaptive, multi-positional agents of teaching. Specifically, teachers may adopt GenAI to refine instructional goals, co-create pedagogical content, negotiate ethical guidelines, and participate in collaborative professional networks. While previous AT-based studies have primarily examined how technologies mediate components within an activity system, the GMAT model contributes this work by theorizing how GenAI contributes to the evolution of sociotechnical and pedagogical ecosystems in L2 education. Furthermore, the study outlines concrete pedagogical implications by examining how specific triadic relationships among the components interact to generate new GenAI-mediated instructional environments. Overall, the GMAT model provides both a theoretical and practical foundation for guiding instructional design and teacher development in the era of GenAI-enhanced education.
- 1 January 2026
- Published by National Foreign Language Resource Center in Language Learning & Technology
- Vol. 30 (1) , 1-21
GenAI-mediated activity theory (GMAT) for L2 teachers
This study proposes the GenAI-Mediated Activity Theory (GMAT) as a conceptual framework for understanding how generative AI (GenAI) reshapes second language (L2) teachers’ teaching preparation and practices within complex pedagogical ecosystems. Methodologically, the paper adopts a model-building approach. It recontextualizes and synthesizes conceptual elements from Engeström’s (1999) expanded Activity Theory (AT) to construct the model. Thus, the study illuminates the evolving role of L2 teachers as adaptive, multi-positional agents of teaching. Specifically, teachers may adopt GenAI to refine instructional goals, co-create pedagogical content, negotiate ethical guidelines, and participate in collaborative professional networks. While previous AT-based studies have primarily examined how technologies mediate components within an activity system, the GMAT model contributes this work by theorizing how GenAI contributes to the evolution of sociotechnical and pedagogical ecosystems in L2 education. Furthermore, the study outlines concrete pedagogical implications by examining how specific triadic relationships among the components interact to generate new GenAI-mediated instructional environments. Overall, the GMAT model provides both a theoretical and practical foundation for guiding instructional design and teacher development in the era of GenAI-enhanced education.
- 25 December 2025
- Published by Oxford University Press (OUP) in ELT Journal
- Vol. 80 (1) , 164-173
No-code approaches to developing ELT resources with generative artificial intelligence
In this series, we explore technology-related themes and topics. The series aims to discuss and demystify what may be new areas for some readers and to consider their relevance for English language teachers. Advances in the no-code approach that uses generative artificial intelligence (GenAI) have enabled practitioners to develop ELT resources and tools without having programming expertise. More recently, within the ELT field, vibe coding has emerged. This approach allows instructors to develop and revise customized apps by interacting with GenAI and by using natural language alone. These innovations hold significant promise for ELT because teachers can now create resources such as chatbots for language practice, automated essay-scoring tools, and language exercise apps tailored to their needs, without relying on IT professionals. This feature article introduces three currently available approaches to developing a GenAI-mediated tool and discusses their merits and challenges in the ELT domain.
- 25 December 2025
- Published by Oxford University Press (OUP) in ELT Journal
- Vol. 80 (1) , 164-173
No-code approaches to developing ELT resources with generative artificial intelligence
In this series, we explore technology-related themes and topics. The series aims to discuss and demystify what may be new areas for some readers and to consider their relevance for English language teachers. Advances in the no-code approach that uses generative artificial intelligence (GenAI) have enabled practitioners to develop ELT resources and tools without having programming expertise. More recently, within the ELT field, vibe coding has emerged. This approach allows instructors to develop and revise customized apps by interacting with GenAI and by using natural language alone. These innovations hold significant promise for ELT because teachers can now create resources such as chatbots for language practice, automated essay-scoring tools, and language exercise apps tailored to their needs, without relying on IT professionals. This feature article introduces three currently available approaches to developing a GenAI-mediated tool and discusses their merits and challenges in the ELT domain.
- Open Access
- 18 November 2025
- Published by Walter de Gruyter GmbH in International Review of Applied Linguistics in Language Teaching
- Vol. 64 (3) , 1589-1615
Exploring prompt iteration process through generative AI image creation in EFL vocabulary learning
Amid the growing use of GenAI in second language education and its emerging multimodality, this study investigated how Korean undergraduates studying EFL enacted prompt iteration with a text-to-image generator to support multimodal vocabulary learning. Participants read etymology-rich word stories, sketched visualizations, and iteratively refined prompts until the images aligned with their intentions. Data included AI-generated artworks, reflective papers documenting the process, and a vocabulary learning strategies survey. Overall, learners reported high use of strategies after the task. Analysis of artworks and reflections revealed a metacognitive–strategic cycle: learners monitored and evaluated outputs, then strategically varied contextual features to negotiate meaning and achieve nuanced imagery. Framed as both a prompt-engineering technique and a literacy practice involving planning, monitoring, and evaluation, prompt iteration emerged as central to GenAI-mediated vocabulary learning, highlighting learners’ agency in exploring word meanings.
- 18 November 2025
- Published by Walter de Gruyter GmbH in International Review of Applied Linguistics in Language Teaching
- Vol. 64 (3) , 1589-1615
Open AccessExploring prompt iteration process through generative AI image creation in EFL vocabulary learning
Amid the growing use of GenAI in second language education and its emerging multimodality, this study investigated how Korean undergraduates studying EFL enacted prompt iteration with a text-to-image generator to support multimodal vocabulary learning. Participants read etymology-rich word stories, sketched visualizations, and iteratively refined prompts until the images aligned with their intentions. Data included AI-generated artworks, reflective papers documenting the process, and a vocabulary learning strategies survey. Overall, learners reported high use of strategies after the task. Analysis of artworks and reflections revealed a metacognitive–strategic cycle: learners monitored and evaluated outputs, then strategically varied contextual features to negotiate meaning and achieve nuanced imagery. Framed as both a prompt-engineering technique and a literacy practice involving planning, monitoring, and evaluation, prompt iteration emerged as central to GenAI-mediated vocabulary learning, highlighting learners’ agency in exploring word meanings.
- 4 November 2025
- Published by Wiley in International Journal of Applied Linguistics
- Vol. 36 (4) , 3086-3097
Translanguaging and Motivational Development in CLIL: A Case From Kazakhstan's Trilingual Context
This study explores how translanguaging is enacted in trilingual content and language integrated learning (CLIL) classrooms in Kazakhstan, and how it shapes students’ motivation to learn English. While translanguaging has gained attention as an analytical lens for examining teachers’ and students’ language practices, little is known about its implementation and motivational impact in contexts where Kazakh, Russian, and English coexist. Drawing on classroom video data and student surveys, this study investigates (1) how teachers and students mobilize multilingual resources to support comprehension; and (2) how students perceive translanguaging versus English‐only instruction in relation to their motivational self‐system. Thematic analysis of classroom interactions revealed that teachers engaged in translanguaging practices, introducing disciplinary concepts, supporting meaning‐making through gestures and realia, and co‐constructing understanding through student‐initiated exchanges. Quantitative findings showed that students strongly favored multilingual approaches over target language (TL)‐only policies, particularly valuing translanguaging for comprehension and peer collaboration. Correlation analyses indicated that learners with strong ought‐to second language (L2) selves preferred first language (L1) support, whereas those with strong ideal L2 selves favored TL immersion. Findings suggest that translanguaging supports both cognitive access and motivational engagement, reinforcing the need for a flexible, multilingual pedagogy tailored to students’ linguistic and affective needs in post‐Soviet CLIL contexts.
- 4 November 2025
- Published by Wiley in International Journal of Applied Linguistics
- Vol. 36 (4) , 3086-3097
Translanguaging and Motivational Development in CLIL: A Case From Kazakhstan's Trilingual Context
This study explores how translanguaging is enacted in trilingual content and language integrated learning (CLIL) classrooms in Kazakhstan, and how it shapes students’ motivation to learn English. While translanguaging has gained attention as an analytical lens for examining teachers’ and students’ language practices, little is known about its implementation and motivational impact in contexts where Kazakh, Russian, and English coexist. Drawing on classroom video data and student surveys, this study investigates (1) how teachers and students mobilize multilingual resources to support comprehension; and (2) how students perceive translanguaging versus English‐only instruction in relation to their motivational self‐system. Thematic analysis of classroom interactions revealed that teachers engaged in translanguaging practices, introducing disciplinary concepts, supporting meaning‐making through gestures and realia, and co‐constructing understanding through student‐initiated exchanges. Quantitative findings showed that students strongly favored multilingual approaches over target language (TL)‐only policies, particularly valuing translanguaging for comprehension and peer collaboration. Correlation analyses indicated that learners with strong ought‐to second language (L2) selves preferred first language (L1) support, whereas those with strong ideal L2 selves favored TL immersion. Findings suggest that translanguaging supports both cognitive access and motivational engagement, reinforcing the need for a flexible, multilingual pedagogy tailored to students’ linguistic and affective needs in post‐Soviet CLIL contexts.