Developing TOPIK Speaking Practice Model and Exploring Effectiveness Using ChatGPT as a Learning Feedback Tool

This study uses ChatGPT, a generative AI, as a learning feedback tool to develop a practice model and customized feedback prompt to improve learners’ TOPIK speaking ability, and to verify its effectiveness through 8 educational experiments. As a research method, the pre-evaluation and post-evaluation results of 10 experimental groups and 10 control groups that did not apply the practice model were compared and analyzed, targeting 20 Chinese learners at the middle and high level based on TOPIK. To this end, the analysis was conducted in a quantitative and qualitative way through the corresponding sample t-test and interview. As a result, the experimental group showed a remarkably more pronounced improvement effect than the control group in the pre-evaluation and post-evaluation results, and significantly higher improvement scores than the control group in three evaluation factors: content and task performance, language use, and speech delivery power. In addition, even in the post-interview of the experimental group, the participants of the experiment showed a positive response to this practice model. This study proved that the use of feedback functions of AI tools such as ChatGPT in the TOPIK speaking education field can substantially help Korean learners improve TOPIK speaking learning ability and is meaningful in showing the potential of AI-based learning tools in TOPIK speaking learning.(Seoul National University)