首页|期刊导航|远程教育杂志|人智协同模式的合规与共治:教育可解释人工智能治理框架构建

人智协同模式的合规与共治:教育可解释人工智能治理框架构建OA

Compliance and Co-governance in Human-AI Collaboration:Constructing a Governance Framework for Explainable Artificial Intelligence in Education

中文摘要英文摘要

生成式人工智能深度融入教育核心环节,在赋能教学革新的同时,其固有的"黑箱"特性也引发了透明度缺失、算法偏见与问责困难等严峻治理挑战,对人智协同教育的实现构成根本障碍.为应对上述挑战,研究旨在构建一个面向人智协同、融贯技术可行性与教育可接受性的教育可解释人工智能综合治理框架.首先,通过批判性整合国际政策与学术理论,廓清教育可解释性的核心概念体系,为治理实践奠定理论基础.其次,从"政策合规—协同治理—能力建设"三个维度构建治理框架:系统解析了以《人工智能法案》为核心的欧盟数字法律生态,并将其转化为适用于教育高风险场景的合规操作清单与治理工具;通过对智能阅卷、课堂行为分析等本土典型案例的深度剖析,揭示了多元利益相关者的差异化解释需求、治理干预与责任共担机制;借鉴国际能力框架,设计了分阶段、分角色的教育者可解释人工智能能力矩阵.最终,提出了从宏观制度到微观实践、从主体赋能到生态培育的系统性路径,不仅为破解教育人工智能的"黑箱"困境提供了系统的理论分析框架,也为在中国教育语境下构建"以人为本、技术向善"的治理新生态提供了可操作性的实践路线.

Generative artificial intelligence(GenAI)is being deeply integrated into core educational processes.While enabling pedagogical innovation,its inherent"black-box"characteristics also generate serious governance challenges,including lack of trans-parency,algorithmic bias,and difficulties in accountability,thereby posing fundamental barriers to human-AI collaborative education.To address these challenges,this study aims to construct a comprehensive governance framework for explainable artificial intelligence in education that is oriented toward human-AI collaboration and integrates technical feasibility with educational acceptability.First,through a critical synthesis of international policy documents and relevant academic theories,the study clarifies the core conceptual system of explainability in education and lays a theoretical foundation for governance practice.Second,a governance framework is de-veloped along three dimensions—policy compliance,collaborative governance,and capacity building.The study systematically inter-prets regulatory compliance requirements exemplified by the EU Artificial Intelligence Act and translates them into actionable compli-ance checklists and governance tools suitable for high-risk educational scenarios.Through in-depth analyses of local representative cases,such as AI-assisted scoring and classroom behavior analytics,the study further identifies stakeholders'differentiated needs for explanation,approaches to governance intervention,and mechanisms for shared responsibility.Drawing on international competency frameworks,it then designs a phased,role-specific competency matrix for educators'explainable-AI capabilities,together with peda-gogical integration pathways.Finally,the study map the landscape spanning macro-level institutions to micro-level practices,and ex-tending from empowering key actors to cultivating an enabling ecosystem.The proposed framework not only offers a structured theoret-ical lens for addressing the"black-box"dilemma of educational AI,but also provides an actionable roadmap for building a human-centered,technology-for-good governance ecosystem within the Chinese educational context.

兰国帅;郑明扬;蒋顷烁;肖琪;宋帆

河南大学教育学部(河南开封 475004)河南大学教育学部(河南开封 475004)河南大学教育学部(河南开封 475004)河南大学教育学部(河南开封 475004)河南大学教育学部(河南开封 475004)

社会科学

教育可解释人工智能治理框架人智协同算法治理教育人工智能教师数字素养

Explainable AI in educationGovernance frameworkHuman-AI collaborationAlgorithmic governanceEducational AITeacher digital literacy

《远程教育杂志》 2026 (1)

51-60,82,11

本文系国家社会科学基金2025年度教育学一般项目"人智协同赋能大规模因材施教的有效模式与推进策略研究"(项目编号:BCA250070)的阶段性研究成果.

10.15881/j.cnki.cn33-1304/g4.2026.01.006

评论