Exploring Student Engagement in AI-Assisted CLIL for Global Citizenship and Competence in EFL Higher Education

This study investigates the impact of an AI-assisted Content and Language Integrated Learning (CLIL) model, designed around Global Citizenship and Competence (GCC), on EFL tertiary students’ motivation and self-efficacy, utilizing goal self-concordance as a key measure of engagement. A 15-week longitudinal study involved 154 South Korean undergraduates enrolled in a required general education course. The course integrated generative AI tools (ChatGPT and DeepL) as translanguaging scaffolds to facilitate engagement with complex global content. Self-efficacy in Language Skills (LS) and GCC were measured pre- and post-course, alongside engagement variables (self-concordance, effort, and perceived outcome) tracked at three time points. Results showed significant increases in self-efficacy for both LS (Hedges’ g = 1.21) and GCC (g = 1.27). Engagement, particularly self-concordance in GCC and digital literacy, also significantly increased. Hierarchical regression analyses revealed that post-course LS self-efficacy was significantly predicted by perceived outcome in digital literacy in GCC and effort in GCC as well as engagement variables in language learning. Post-course GCC self-efficacy was strongly predicted by self-concordance, effort, and perceived outcome specific to the GCC domain. These findings support the premise that AI-assisted CLIL cultivates self-efficacy by fostering engagement in higher-order learning processes, highlighting the reciprocal nature of language and content engagement within this inclusive pedagogical framework.