Fostering sustainable cross-cultural learning through AI-supported multimodal storytelling activities

Cross-cultural learning is increasingly important in higher education, yet the use of multimodal generative artificial intelligence (AI) and Retrieval-Augmented Generation (RAG) to support authentic cross-cultural learning remains under-examined. This study developed a web-based system integrating multimodal generative AI with RAG technology, aimed at enhancing learners’ Multicultural Competence (MC), Critical Intercultural Awareness (CIA), Ethnocultural Empathy (EE), and sustainable cross-cultural learning experiences. The system combined cultural video creation, scaffolded essay writing, AI feedback, chatbot assistance, and peer interaction. A quasi-experimental design involving 42 students compared an Experimental Group (EG), which used the developed system, with a Control Group (CG), which used general AI-supported writing tools to complete the same tasks. Learning activities focused on tea ceremonies and traditional costumes and followed a four-step model emphasizing cultural transformation and sustainability. Data were collected through pre- and post-questionnaires, participant perception, cross-cultural knowledge tests, essays, and interviews. Quantitative results found that the EG showed significantly greater improvements in MC, CIA, and EE than the CG and reported high satisfaction, while qualitative results further corroborated these findings. The study contributes to theory by expanding the framework of multicultural competence and to practice by demonstrating the application of a RAG-integrated multimodal AI system in cross-cultural education.