«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

清华大学教育研究[ISSN:1001-4519/CN:11-1610/G4]

卷:

期数:

2026年05期

页码:

29-39

栏目:

人工智能与教育

出版日期:

2026-10-20

文章信息/Info

Title:

The Generative Turn in the Structure of Educational Objectives in the Age of Arti?cial Intelligence: Limitations of Bloom’s Taxonomy and a Way Beyond

作者:

肖桐¹ 刘柳²

1.厦门大学 教育研究院；2.新加坡国立大学 人工智能教育技术中心

Author(s):

XIAO Tong¹ LIU Liu²

1. Institute ofEducation, Xiamen University;2. Center for Arti?cial Intelligence in Education Technology, National University of Singapore

关键词:

生成式人工智能; 布鲁姆分类学; 生成型目标结构; 存在伦理

Keywords:

generative arti?cial intelligence;  Bloom’s taxonomy;  generative objective structure;  existential ethics

分类号:

G434

DOI:

10.14138/j.1001-4519.2026.05.002911

文献标志码:

A

摘要:

在生成式人工智能深度介入认知活动与教育制度的背景下，布鲁姆教育目标分类体系所依赖的线性递进、功能分解与行为映射逻辑，正因认知路径网络化、教学主体协同化与评价过程生成化而显现出新的结构张力。其问题已触及教育目标如何获得表达形式、组织结构与正当性基础这一更为根本的问题。本文在分析布鲁姆分类体系认知逻辑、制度嵌入及其在AI时代适用边界的基础上，提出“生成型目标结构”。该结构由生成语法、生成图谱与生成合法性三个相互关联的维度构成。生成语法赋予目标以过程性、关系性与开放性，生成图谱以多节点及其动态联系重组目标之间的结构关系，生成合法性则将回应性、未竟性与主体生成纳入目标正当性的判断基础。三者共同推动教育目标由固定表述与层级排序走向动态调适与关系生成，使目标在保持方向性与规范性的同时，为学习路径变化、意义生成与主体展开保留开放空间。生成型目标结构重新界定了布鲁姆分类体系在AI认知生态中的适用边界，并为课程设计、教学实践与教育治理提供一种理解复杂学习过程及其伦理正当性的新框架。

Abstract:

As generative artificial intelligence becomes deeply embedded in cognition and educational systems, the logics underlying Bloom’s taxonomy—linear progression, functional decomposition, and behavior-based mapping—are encountering new structural tensions arising from networked cognitive pathways, collaboration among participants in teaching and learning, and increasingly generative forms of assessment. The challenge now extends to more fundamental questions concerning how educational objectives are expressed, organized, and justi?ed. Building on an analysis of the cognitive logic, institutional embedding,and contemporary limits of Bloom’s taxonomy, this article proposes a generative objective structure composed of three interrelated dimensions: generative grammar, generative map, and generative legitimacy.Generative grammar gives educational objectives greater processual, relational, and open-ended qualities; the generative map reorganizes their structural relations through multiple nodes and dynamic connections; and generative legitimacy incorporates responsiveness, non-finality, and subject formation into the grounds on which educational objectives are justi?ed. Together, these dimensions shift educational objectives from ?xed formulations and hierarchical ordering toward dynamic adjustment and relational emergence, while maintaining their directional and normative functions and leaving room for changes in learning pathways, meaning-making,and subject development. The proposed structure thus rede?nes the limits of Bloom’s taxonomy within an AI-mediated cognitive ecology and offers a new framework for understanding complex learning processes and their ethical legitimacy in curriculum design, teaching practice, and educational governance.

常用功能

更新日期/Last Update: 2026-10-20