ReefPaperAgent: Empowering Teaching Case Generation with Large Language Models

Facing the challenge of obtaining negative cases for critical pedagogy in traditional academic writing courses, this paper proposes a methodology that uses large language models (LLM) to generate customized negative cases for academic writing. According to students' research interests and typical errors in academic writing, a negative cases generation workflow is given, including data feeding, role definition, response logic setting, prompts design, and workflow development. Then, an AI agent named 'ReefPaperAgent' is developed on 'Coze' platform, and this AI agent can generate negative cases for academic writing automatically. Finally, using these negative cases in critical pedagogy, our academic writing course effectively attracts students.