中国电化教育

随着生成式人工智能等智能技术的蓬勃发展,人智协同问题解决正逐渐成为未来工作情境的新常态。然而在教育领域,当前研究仍缺乏对人智协同问题解决的概念内涵、过程模型及其运行机制的深入探讨。研究从混合智能的视角对人智协同问题解决进行了明确的概念界定,并剖析了其主要特征,包括能力结构的互补性、协同方式的多样性以及认知发展的同生性。在此基础上,首先构建了人智协同问题解决的过程模型,分析了模型的核心要素;其次,依托协同理论、“人在回路”理论和社会认知理论,探讨了其运行机制。组织机制以协同理论为基础,通过角色分工、信息共享与过程协调实现开放任务情境中的结构化推进;动力机制关注通过目标、过程与价值的多层对齐实现动态同步,确保协作过程的可控性与教育导向;交互机制基于社会认知理论,通过示范、模仿、探究与反馈的循环,促进策略迁移、知识建构与协同探究。将人智协同问题解决过程模型应用于面向问题解决的教学,以设计性问题为例,设计并开发了人智协同问题解决多智能体系统,分析其在真实教学情境中的适配性与可操作性。

With the vigorous development of intelligent technologies, such as generative artificial intelligence, human-AI collaborative problemsolving is gradually becoming the new norm. However, in the field of education, current research still lacks in-depth examination of the conceptual connotation, process models, and operational mechanisms of human-AI collaborative problem solving. From the perspective of hybrid intelligence, this study provides a clear conceptual definition of human-AI collaborative problem solving and analyzes its main characteristics, including the complementarity of ability structures, the diversity of collaboration modes, and the co-developmental nature of cognition. Building on this, the study constructs a process model of human-AI collaborative problem-solving and analyzes the core elements of the model. Second, drawing on the “Human-in-the-Loop” theory, collaboration theory, and social cognitive theory, it systematically explores its operation mechanisms. The organizational mechanism, grounded in collaboration theory, enables structured progression in open task contexts through role distribution, information sharing, and process coordination; the motivational mechanism centers on attaining dynamic synchronization via multi-layered alignment of goals, processes, and values, ensuring both the control ability of the collaborative process and its adherence to educational purposes; and the interactive mechanism, informed by social cognitive theory, facilitates strategy transfer, knowledge construction, and collaborative inquiry through cycles of modeling, imitation, exploration, and feedback. The human-AI collaborative problem-solving process model was implemented in problem-oriented instructional design. Using design-based problem tasks as an exemplar, a multi-agent system supporting human-AI collaborative problem solving was designed and developed, and its pedagogical alignment and operational feasibility were analyzed in instructional context.