AI-enhanced cross-cultural learning with contextualization, personalization, and socialization in authentic contexts

Connecting cross-cultural learning with authentic real-life contexts remains challenging, despite its importance for developing learners’ cross-cultural competencies. This study developed and implemented an artificial intelligence (AI)-based X-Cultural App to support cross-cultural learning in authentic environments. The application integrates six functions: authentic context recognition, sample sentence generation, scaffolded essay writing, a personalized question generation bank, real-time AI feedback, and Q&A support. By leveraging multiple image-to-text (ITRs) recognition technologies and Generative Pre-trained Transformer (GPT) models, the system provides contextualized input, adaptive scaffolding, and interactive feedback to enhance learning. Two AI-supported cross-cultural learning activities focusing on traditional cuisine and costumes were designed based on the Contextualization–Personalization–Socialization (CPS) model. The study involved 25 graduate students from Chinese, Thai, Indonesian, and Vietnamese ethnicities and examined the development of intercultural communicative competence (ICC) and intercultural sensitivity (ISS). Multiple data sources, including pre- and post-tests, essays, questionnaires, and interviews, were collected for triangulation analysis. Results showed significant improvements in participants’ ICC and ISS, supported by qualitative evidence from essays and interviews. Participants also reported positive perceptions of the AI-based X-Cultural App, particularly regarding its usefulness, ease of use, and engagement. The findings provide practical insights for designing AI-supported cross-cultural learning environments.