<![CDATA[Background. The rapid advancement of artificial intelligence has transformed language learning environments, particularly through the integration of Large Language Models (LLMs) that facilitate interactive and personalized communication practice. Despite their growing adoption in English as a Foreign Language (EFL) education, limited empirical attention has been devoted to understanding their influence on learners’ pragmatic competence and speaking anxiety. Purpose. This study aims to evaluate the impact of AI-powered dialogue systems based on LLMs on EFL learners’ ability to use language appropriately in social contexts and their confidence in oral communication. Method. A quantitative quasi-experimental design was employed involving 120 EFL learners divided into experimental and control groups. Data were collected through pragmatic competence assessments, speaking anxiety scales, classroom observations, and structured interviews. Results. Findings indicate that students who engaged with AI-powered dialogue systems demonstrated significantly higher gains in pragmatic competence compared to those receiving conventional instruction. Results also reveal a substantial reduction in speaking anxiety, attributed to the non-judgmental, adaptive, and accessible nature of AI-mediated interactions. Conclusion. The study concludes that LLM-based dialogue platforms provide meaningful opportunities for authentic communication practice while fostering psychological comfort in language learning. Integrating AI-powered dialogue into EFL instruction can therefore enhance communicative effectiveness and support learners’ confidence in real-world language use.]]>
Copyrights © 2026
