远程教育杂志2025，Vol.43Issue(6)：52-62,11.DOI:10.15881/j.cnki.cn33-1304/g4.2025.06.007
大模型多智能体支持下人机群智协同教研模式的构建与验证
Construction and Validation of a Human-Machine Collaborative Teaching and Research Model Supported by Large Language Model-Based Multi-Agent
摘要
Abstract
In the era of large language models,human-machine collaborative teaching and research has become an increasing-ly important approach to supporting teacher professional development.However,existing collaborative practices frequently rely on a single intelligent agent with limited and homogeneous functions,which may lead to an imbalance in human-machine role distribution and reduce the effectiveness of collaborative inquiry.Consequently,such practices constrain the generation and resolution of cognitive conflict among participants and impede the development of deep teacher reflection.Guided by collective intelligence theory and expe-riential learning theory,this study investigates the operational mechanisms underlying human-machine collaborative teaching and re-search.On this basis,the study develops a multi-agent system using platforms such as Feishu,COZE,DeepSeek,and Doubao,and constructs a practical model to support human-machine collaborative teaching and research.To evaluate the effectiveness of this mod-el,a quasi-experimental study was conducted with 40 primary and secondary school teachers.The findings indicate that,compared with collaboration involving a single intelligent agent,the multi-agent system more effectively enhances the quality of humammachine interaction,significantly increases teachers'levels of group cognitive conflict,and promotes deeper reflective engagement.Future re-search should focus on enhancing multidimensional interactions within collective-intelligence systems,improving the coherence and comprehensiveness of role allocation,and exploring mechanisms that optimize human-machine collaboration,thereby contributing to the high-quality development of teaching and research in the intelligent era.关键词
大语言模型/多智能体/群体智慧/教师教研/人机协同Key words
Large language models/Multi-agent/Collective intelligence/Teacher teaching and research/Human-machine collaboration分类
社会科学引用本文复制引用
李典洋,李海峰..大模型多智能体支持下人机群智协同教研模式的构建与验证[J].远程教育杂志,2025,43(6):52-62,11.基金项目
本文系2024年度国家社会科学基金教育学西部项目"大语言模型会话机器人支持的西北边疆乡村教师高质量教研研究"(项目编号:XCA240326)、新疆师范大学研究生科研创新项目"面向深度教学的多智能体群智协同模式构建研究"(项目编号:XSY202501003)、新疆师范大学研究生科研创新项目"走出元认知惰性泥潭的人智协同调节模式研究"(项目编号:XSY202501029)的研究成果. （项目编号:XCA240326）
