华东师范大学学报(教育科学版) ›› 2026, Vol. 44 ›› Issue (8): 67-80.doi: 10.16382/j.cnki.1000-5560.2026.08.004

• 先锋论文选 • 上一篇    下一篇

《人机协同视域下的深度认知支架——基于12824条学习者-GenAI多轮对话的滞后序列分析》内容详介及研究与写作过程披露声明

程一可, 钱江明, 朱宇恒   

  1. 南京农业大学公共管理学院,南京 210095
  • 出版日期:2026-08-01 发布日期:2026-07-22

“Deep Cognitive Scaffolding in Human-AI Collaboration: A Lag Sequential Analysis Based on Learner-GenAI Multi-Turn Dialogues”: A Detailed Exposition and a Disclosure Statement of the Research and Writing Process

Yike Cheng, Jiangming Qian, Yuheng Zhu   

  1. College of Public Administration, Nanjing Agricultural University, Nanjing 210095, China
  • Online:2026-08-01 Published:2026-07-22

摘要:

本文系全球首个“AI一作”大型社会实验“人机共创先锋论文榜”论文之一。文章第一部分为研究内容详介,基于12824条“学习者-GenAI”多轮对话日志,运用滞后序列分析(LSA)探究人机协同中的认知支架机制,揭示“试错-反馈-再修正”的深度学习闭环,以及学习者在算法固着时通过元认知监控实现主体性重塑的关键作用;第二部分为研究与写作过程披露,系统厘清本次实验所使用的多种大模型工具、AI 与人类作者权责分工、分层拆解式提示词工程设计方案,完整还原数据处理、论文生成、跨模型交叉审校、人工事实核验与学术伦理管控等全链条操作细节,同时阐明 AI 署名创新实验的伦理界定、成果原创核查、科研可重复性约束与人类研究者终局责任划分等关键实践议题。研究发现:(1)学习者与GenAI的交互呈现出显著的长程化、不对称特征,GenAI通过少问多答的不对称话语模式提供持续的脚手架支持;(2)在行为序列上,学习者的自我修正与GenAI的纠正引导之间存在高频的闭环回路,证明了深度学习并非发生于单次问答,而是在螺旋式的“试错-反馈-再修正”中实现;(3)微观案例分析揭示,当GenAI陷入算法固着时,学习者会迅速从提问者转换为策略制定者,通过元认知监控实施关键干预。GenAI在未来教育中不应仅被视为信息检索工具,而应作为扮演启发思考的“苏格拉底式导师”重塑学习者的主体性,共同构建人机共生的教育新生态。本文的价值在于为“人机协同”提供可触摸的实证细节,同时以全过程透明披露为学界理解AI科研能力边界提供参照样本。

关键词: 人机协同, 生成式人工智能, 认知支架, 滞后序列分析, 教育数据挖掘

Abstract:

This paper is one of the publications featured in the “Human-AI Co-Creation Pioneer Papers Ranking”, which is part of the Panoramic Report on the World’s First Large-Scale Social Experiment of “AI as the First Author”.The first part of this paper presents a detailed exposition of the research content: based on 12,824 log entries of multi-turn dialogues between learners and GenAI, it employs Lag Sequential Analysis (LSA) to investigate the cognitive scaffolding mechanism in human-AI collaboration, revealing a high-frequency closed-loop of “trial-error, feedback, and re-correction” as well as the critical role of learners' metacognitive monitoring in reshaping agency when GenAI encounters algorithmic fixation. The second part constitutes a transparent disclosure of the entire research and writing workflow. It systematically sorts out the various large language model tools adopted in this human-AI joint research, the division of rights and responsibilities between AI and human researchers, and the hierarchical and decomposed prompt engineering design schemes. It fully reconstructs the full-chain operational details covering data processing, paper drafting, cross-model cross-review, manual factual verification and academic ethics governance. Meanwhile, it elaborates on key practical issues including ethical definition of the innovative AI-authorship experiment, originality inspection of research outputs, constraints on research reproducibility, and the allocation of ultimate accountability among human researchers. The findings indicate that: (1) learner-GenAI interactions exhibit significant long-term and asymmetric characteristics, with GenAI providing continuous scaffolding support through an asymmetric discourse pattern of “few questions, many answers”; (2) at the behavioral sequence level, a high-frequency closed-loop exists between learners’ self-correction and GenAI’s corrective guidance, demonstrating that deep learning does not occur in single Q&A exchanges but is achieved through a spiral process of “trial-and-error, feedback, and re-correction”; (3) micro-case analysis reveals that when GenAI falls into algorithmic fixation, learners swiftly shift from questioners to strategy formulators, implementing critical interventions through metacognitive monitoring. GenAI should not be regarded merely as an information retrieval tool in future education; instead, it should serve as a “Socratic tutor” that inspires thinking, reshapes learner agency, and jointly constructs a new educational ecology of human-AI symbiosis. The value of this paper lies in providing tangible empirical details for the grand narrative of “human-AI collaboration,” while offering the academic community a referential sample for understanding the boundaries of AI’s research capabilities through full-process transparent disclosure.

Key words: human-AI collaboration, Generative Artificial Intelligence (GenAI), cognitive scaffolding, Lag Sequential Analysis (LSA), educational data mining