The rapid advancement of large language models (LLMs) has created new possibilities for automating knowledge delivery in higher education. While existing research has explored AI-assisted content generation, little attention has been paid to the structured collaboration between human educators and LLMs in the process of instructional material creation. This paper proposes a Human-LLM Collaborative Framework (HLCF) for the automatic generation of university course slides. The framework integrates human pedagogical expertise and LLM generative capabilities across four stages: syllabus definition, automated draft generation, expert revision, and consistency optimization. To evaluate the framework, a case study was conducted in an undergraduate "Artificial Intelligence and Society" course. Quantitative and qualitative analyses revealed that the HLCF reduced slide preparation time by more than 60% while maintaining or improving pedagogical coherence and content quality. The study also identifies key human roles—such as prompt engineering, content validation, and contextual adaptation—that remain essential in ensuring academic rigor and alignment with intended learning outcomes. The findings highlight the potential of human-AI collaboration to enhance instructional design efficiency, bridge the gap between automation and academic integrity, and contribute to the evolving paradigm of intelligent education systems.
Xiaoming et al. (2025) studied this question.