Designing and Evaluating a Voice-Based Chatbot System for Second Language Speaking Fluency and Learner Engagement

This pilot study investigates the potential of a voice-based chatbot (EnMIA) to support speaking fluency, motivation, and engagement among undergraduate Korean language learners at a single U.S. Midwestern university. The chatbot aims to provide interactive, real-world speaking tasks accessible through multi-platforms, supporting seamless learning. Data were collected over one academic semester through pre- and post-speaking assessments and surveys. Speaking performance data indicated an improvement in fluency, though accuracy and complexity remained unchanged, in a pre-post design without control group, suggesting short-term practice may strengthen learners’ ability to speak more smoothly and confidently. However, without a control group, gains cannot be solely attributed to the tool. Survey results showed high perceptions of support, design, and usability, with interactive tasks correlated with motivation. The findings highlight chatbot-supported interaction potentially enhance motivation and fluency, while pointing to the need for extended practice and targeted design to affect accuracy and complexity.