Transforming English education in Japan by utilizing large language models for tailored learning and diverse skill development

Abstract

The integration of advanced generative AI has profoundly influenced higher education, especially in language learning. In Japan’s exam-focused and teacher-centered English education system, students have few opportunities to use generative AI to enhance critical skills like analysis, evaluation, and creativity. Learning often remains passive, with an emphasis on memorization and testing rather than on higher-order thinking skills (HOTS). To address these issues, the authors propose a new AI-driven approach to language teaching that combines HOTS with lower-order thinking skills (LOTS), based on Bloom’s revised taxonomy. This method aims to promote deeper engagement, critical thinking, and active learning through AI-assisted activities such as information synthesis and problem-solving. The approach utilizes large language models, like LLaMA 3, to offer personalized learning experiences tailored to students’ English abilities and individual interests. Pre-prompts guide learning toward specific objectives such as vocabulary, grammar, and pronunciation. Furthermore, the plan includes integrating the system with an AI assistant robot to increase interactivity and immersion. While concerns about diminished human interaction, privacy, algorithmic bias, and accessibility persist, AI-powered language learning has the potential to bring significant advantages to Japan’s higher education system. By providing a supportive, non-judgmental environment, generative AI can help overcome the challenge of communication in English, boost student confidence, and support the digital transformation of education, ultimately aiding students in preparing to live in a global society.