Exploring Students’ Experiences with Generative AI in Non-English Higher Education Landscape: Case Studies from Japan and Mongolia

• In Event: Highlighted Session: Perils and Promises of Generative AI in Higher Education: Comparative Insights from Mainland China, Hong Kong, Japan, Mongolia, and the U.S.

Proposal

Generative AI (GAI) introduces both profound opportunities and considerable challenges for university students, particularly within non-English speaking regions. Predominantly optimized for English text processing, GAI has limitations in handling non-English languages which may lead to varied experiences influenced by linguistic, cultural, and societal factors. This study aims to illustrate some of these challenges faced by students in Japan and Mongolia when using GAI.

This study draws on the Unified Theory of Acceptance and Use of Technology (UTAUT) by Venkatesh (2016), which provides a comprehensive framework to explore the factors that influence students’ acceptance and use of GAI in higher education. UTAUT considers key determinants such as performance expectancy, effort expectancy, social influence, and facilitating conditions. By applying this model, the study examines how these factors, alongside linguistic and cultural contexts, shape students’ experiences with GAI. The framework’s emphasis on both individual and contextual factors offers valuable insights into the unique challenges and opportunities GAI presents in non-English speaking educational environments.

A qualitative approach was chosen for its effectiveness in exploring students’ personal experiences and the context of GAI usage. We conducted interviews with undergraduate students from one top-tier university in each country, posing two primary research questions: 1) How do students utilize GAI, and what challenges do they encounter? 2) What are the comparative experiences of students from Japan and Mongolia? The sample was methodically balanced in terms of gender and academic discipline to enhance the robustness of the data.

In Mongolia, we interviewed 14 students—5 male and 9 female students across seven schools or faculties. Students came from 6 different provinces or cities. Each interview lasted approximately 30 minutes to one hour. In Japan, the research team will finish interviews in October. The preliminary finding shows that students began using GAI, particularly ChatGPT, as a tool for academic tasks such as essay writing and language learning, often combining it with other tools like Google Translate. However, the accuracy of GAI in processing Mongolian text was a significant challenge, leading some students to prefer English or avoid GAI for complex tasks. Some students reported that GAI helped them become more organized in their study practices by providing quick insights and suggestions, although concerns about its reliability and the risk of dependency on AI-generated content were also highlighted​.

The preliminary results show that Japanese and Mongolian students share common challenges when using GAI, primarily due to the tools’ limited optimization for their respective languages. Both groups struggle with GAI’s handling of complex grammar, formal academic language, and culturally nuanced expressions, often leading to inaccurate or contextually inappropriate outputs. The lack of localized datasets and support for specialized terminology further diminishes the utility of GAI in academic settings.

This study contributes to the ongoing global discourse on the implications of GAI in higher education. Through a qualitative lens, we provide nuanced insights into students’ engagement with GAI, particularly in non-English linguistic contexts. Our findings offer critical implications for the international comparative analysis of GAI’s role in education.

Authors

• Ming Li, Osaka University

• Ariunaa Enkhtur, Osaka University

• Lilan Chen