基于复杂适应系统理论的教育提示语框架构建研究

An Educational Prompt Framework Based on Complex Adaptive Systems Theory

摘要: 生成式人工智能的发展推动了教育提示语研究的兴起，并为人机协同学习提供了新的可能。然而，现有提示语框架主要表现为通用结构型和教学功能导向型两类，在学习过程支持、人机动态交互以及元认知发展等方面仍存在不足。基于复杂适应系统理论，本研究提出CREATE教育提示语框架，包含情境与角色、请求与目标、执行与参数、适应与交互、思维与元认知以及评估与输出六个模块，构建从任务启动到成果评价的人机协同学习闭环。为增强框架的可操作性，研究进一步设计了模块化提示语模板，并提出CREATE-Lite简化版本。通过文献综述写作案例，展示了框架的运行机制及其在教育场景中的应用。研究表明，CREATE框架能够为学习者提供结构化的人机交互支架，支持动态反馈与认知调节，为教育提示语设计提供新的理论视角与实践参考。

Abstract: The rapid development of generative artificial intelligence (GenAI) has stimulated growing interest in educational prompting and created new opportunities for human-AI collaborative learning. However, existing prompting frameworks can generally be categorized as either general structural frameworks or pedagogically task-oriented frameworks, both of which exhibit limitations in supporting learning processes, dynamic human-AI interaction, and metacognitive development. Drawing on Complex Adaptive Systems (CAS) theory, this study proposes the CREATE framework for educational prompting, which consists of six interconnected modules: Context and Role, Request and Target, Execution and Parameters, Adaptation and Interaction, Thinking and Metacognition, and Evaluation and Output. Together, these modules establish a human-AI collaborative learning cycle spanning task initiation, interaction, reflection, and evaluation. To enhance practical applicability, the study further develops a modular prompting template and introduces CREATE-Lite, a simplified version for novice users. A literature review writing scenario is employed to illustrate the operational mechanism and educational application of the framework. The findings suggest that the CREATE framework provides a structured scaffold for human-AI interaction, facilitating dynamic feedback, metacognitive regulation, and deeper learning. This study offers both theoretical insights and practical guidance for educational prompt design in the era of generative AI.