A Generative AI-based Teaching Material System Using a Human-In-The-Loop Model

Teaching materials are crucial for improving the efficiency of teaching and learning. However, traditional tools for creating teaching materials are mutually independent and lack sufficient intelligence, resulting in low efficiency and limited adaptability. This paper presents a generative artificial intelligence(GAI)-based teaching material system(TMGS) to create materials for both teachers and students in a human-in-the-loop model. The TMGS comprises three modules: pre-class material generation, in-class material generation, and post-class material generation. The pre-class module assists teachers in creating well-structured teaching plans, while the in-class module provides valuable resources for both teachers and students during lessons. The post-class module focuses on delivering personalized assignments to students. Throughout all stages, teachers play a critical role in reviewing the generated materials. To evaluate the system’s performance, we conducted materials generation on the topic of “inverse proportion” in primary school mathematics by utilizing large language models like GPT-4. The experiments demonstrate that the system exhibits excellent comprehensiveness, efficiency, and adaptability.