Enhancing EFL Speaking Performance through AI-Powered Conversational Agents: A Mixed-Methods Study at Dong Nai Technology University

Abstract The current paper examines the efficiency of AI-assisted conversational agents in enhancing the speaking proficiency of Vietnamese EFL university students. To conduct the research, a mixed methods design was used, including a quasi-experiment along with questionnaire and interview-based qualitative data. In total, 80 participants were randomly divided into experimental and control groups, during the course of eight to ten weeks. The results show that there is a noticeable improvement in the speaking proficiency of the experimental group in terms of fluency, accuracy, and complexity when compared to those in the control group who have received conventional training. Besides, the former group of learners became more motivated, confident, and less anxious about speaking English. Despite the benefits, some issues such as technical difficulties and inaccurate feedback have been mentioned.