Presentation Information
[W-07]A Design-Based Study of AI-Enhanced Multimodal Support for Content Understanding in an English-Medium History CourseToward Data-Driven Education for Society 5.0
○Ujwal Kumar1, ○Hyoseok Choi1, Hatsuko Yoshikubo1 (1. Shibaura Institute of Technology, College of Engineering)
Keywords:
AI-Assisted Multimodal Learning,AI-Driven Instructional Material Optimization,CEFR-Based Vocabulary Development,EMI Humanities Module in Engineering,Educational Innovation for Society 5.0
This design-based study, a collaborative project with student authors, explores the development of CEFR-level English vocabulary use among international students enrolled in an English-medium instruction (EMI) history module at Shibaura Institute of Technology. Customised lecture slides at two CEFR levels—B2 and C2—were introduced using Claude 3.5 to accommodate the diversity of learners’ English proficiency. Supplementary audio review materials with AI-generated voice narration were provided to enhance comprehension. Student writings were analysed using holistic tools such as the CEFR-based Vocabulary Level Analyser (CVLA) and the AI-powered CEFR-based Writing Level Analyser (CWLA). Initial findings indicate that the mean CEFR score assessed by CVLA rose from 4.65 (just below B2.2) in Week 1 to 5.32 (within the C1–C2 range) in Week 4, with a statistically significant gain (p < 0.01). Final results, completed in July 2025, are expected to inform the design of inclusive, data-driven educational environments aligned with the Society 5.0 vision.