【作 者】：
白雪梅， 霍玉莉， 孙 彧， 顾小清 ， 苏 瑞， 芦嘉玥
【关 键 词 】：
认知临场感； 生成式人工智能； 认知网络分析； 滞后序列分析
【栏 目】：
课程与教学
【中文摘要】：
认知临场感是衡量学生知识建构程度的重要指标，然而其发展仍面临挑战。生成式人工智能（GAI）为促进认知临场感的发展提供了前所未有的机遇。为此，研究通过两项实验探讨GAI对在线协作知识建构中认知临场感发展的促进作用及其影响因素。研究采用认知网络分析法与滞后序列分析法，对在线协作知识建构学习中学生的会话数据进行分析。研究发现：（1）实验组（GAI支持）认知临场感表现优于对照组（无GAI支持），实验组在“解决”阶段的节点强度和“探究—解决”的连接强度均高于对照组。（2）实验组各组认知临场感表现存在差异，具体表现在“触发事件—解决”“探究—解决”和“整合—解决”的连接强度上。（3）共享元认知是影响GAI支持的在线协作知识建构中认知临场感发展的关键，高共享元认知水平组在“探究—解决”连接强度上高于中、低水平组，高、中共享元认知水平组在“整合—解决”连接强度上高于低水平组。（4）不同共享元认知组的认知临场感发展轨迹存在明显差异：高水平组主要集中在整合与解决阶段；中水平组多处于探究阶段；低水平组则频繁触发新问题，且难以从整合过渡到解决。
【英文摘要】：
Cognitive presence is an important indicator of how well students construct knowledge, yet its development continues to face challenges. Generative Artificial Intelligence(GAI) provides an unprecedented opportunity to facilitate the development of cognitive presence. To this end, the study explored the role of GAI in facilitating the development of cognitive presence in online collaborative knowledge construction and its influencing factors through two experiments. The study used cognitive network analysis and lag sequential analysis to analyze students' conversational data in online collaborative knowledge construction learning. It is found that: (1) the experimental group(with GAI support) perform better than the control group(without GAI support) in terms of cognitive presence, and the node strength of the experimental group in the "resolution" phase and the connection strength of the "exploration-resolution" phase are higher than those of the control group. (2) Variations in cognitive presence are observed across experimental subgroups, specifically nanifested in the connection strengths of "triggering event-solution" phase, "exploration-solution" phase and "integration-solution" phase. (3) Shared metacognition is the key to the development of cognitive presence in GAI-supported online collaborative knowledge construction. The high shared metacognition group demostrates a higher connection strength in "exploration-solution" phase than the medium and low shared metacognition groups, while both high and medium shared metacognition groups have a higher connection strength in "integration—resolution" phase than the low shared metacognition group.(4) There are significant differences in the developmental trajectories of cognitive presence between different shared metacognition groups: the high-level group mainly operate in the integration and solution phases; the medium-level group primarily engage in the exploration phase; and the low-level group frequently triggers new problems and struggle to transition from integration to solution.