Learning by Quiz Creation: A Novel Approach Using Interactive Generative AI and Its Experimental Application
Hisaya Kobayashi, Masanori Takagi · 2024
Research has shown that learning through quiz creation, where learners generate their own quizzes, demands a more profound understanding than learning through quiz-solving. Consequently, it is anticipated to yield greater learning outcomes. However, conventional support systems for learning through quiz creation have demonstrated that the learning outcomes vary depending on the process when learners are responsible for creating the quizzes. Additionally, it has been demonstrated that the flexibility of the quiz creation is reduced when the process is streamlined and regulated. To effectively address these issues, we must clarify the quiz creation process to deepen understanding and encourage it in accordance with the process. Moreover, to facilitate the flexible quiz creation and promote a quiz creation process that enhances understanding, we focused on using an interactive generative artificial intelligence (AI). We first analyze the quiz creation process by focusing on three common types: story-based, quiz-based, and solution-based. Additionally, we develop generative AI prompts to facilitate the process. Subsequently, we conducted an experimental application of the generative AI with the designed prompts to university students. The results of the experimental application indicate that the proposed quiz creation process partially contributed to enhancing comprehension through quiz creation, and our approach successfully improved learning outcomes. This study suggests that generative AI can improve learning outcomes by guiding learners to the appropriate process of quiz creation, and that it can help with flexible quiz creation, including the "natural language expression of a quiz setting".