Abstract

Developing speaking competence remains a major challenge in English as a Foreign Language (EFL) instruction at high level. Recent advances in artificial intelligence (AI) offer new possibilities for enhancing speaking instruction through increased practice, interaction, and feedback. This study investigates the effectiveness of AI-supported speaking instruction among undergraduate students at a university in Hanoi. Using a quasi-experimental design, the study involved 70 undergraduates, with 35 assigned to an experimental group and 35 to a control group. The experimental group received AI-assisted speaking activities (chatbots and speech recognition tools) integrated into regular instruction (20-30 minutes per session) in eight weeks, while the control group followed traditional speaking instruction. Speaking performance was assessed through pre- and post-tests and analyzed across fluency, accuracy, complexity, and interactional competence. Independent-samples and paired-samples t-tests were employed for data analysis. The results reveal no significant difference between the two groups at pre-test. However, the experimental group significantly outperformed the control group at post-test, with a large effect size. Notable gains were observed in fluency and interactional competence, significant improvements in accuracy, and more modest gains in complexity. These findings suggest that AI-supported instruction can effectively enhance EFL speaking competence in higher education.

Keywords: Artificial intelligence in language education, EFL speaking instruction, speaking competence, higher education, Vietnamese university context.