VietTESOL International Convention 2026

The integration of artificial intelligence (AI) into Content and Language Integrated Learning (CLIL) has gained increasing attention as an innovative approach to fostering deep learning in English as a Foreign Language (EFL) science classrooms. This study examined the effectiveness of AI-supported CLIL instruction—utilizing AI chatbots and interactive platforms—in improving students’ science learning outcomes, English communication skills, learner autonomy, and classroom engagement in Vietnamese lower secondary education. Grounded in the CLIL framework, TPACK, and Deep Learning Theory, the study employed a quasi-experimental mixed-methods design involving 281 students and 19 science teachers over a 16-week intervention. Participants were divided into an experimental group receiving AI-supported CLIL instruction and a control group receiving conventional instruction. Quantitative data analyzed via SPSS 26 revealed statistically significant improvements in the experimental group’s science achievement and English communication performance compared to the control group (p < .05). Paired-samples t-tests indicated substantial gains in conceptual understanding (M = 68.42 to 84.73), autonomy (M = 3.12 to 4.28), and engagement (M = 3.35 to 4.41). Regression analysis further demonstrated that AI-supported interaction significantly predicted deep learning outcomes (β = .62, p < .001). Qualitative findings confirmed that students became more confident in scientific discussions and demonstrated stronger critical thinking. The study suggests that AI-enhanced CLIL provides an effective pedagogical model for promoting deep learning and technology-integrated science instruction in EFL secondary contexts.
Keywords: Artificial Intelligence; Content and Language Integrated Learning; Deep learning; Mixed-methods; Secondary science education.