华东师范大学学报(教育科学版) ›› 2026, Vol. 44 ›› Issue (9): 70-84.doi: 10.16382/j.cnki.1000-5560.2026.09.007

• 教育数智化转型 • 上一篇    下一篇

教育如何应对生成式人工智能的表征危机

杨澄宇   

  1. 华东师范大学课程与教学研究所,上海 200062
  • 出版日期:2026-09-01 发布日期:2026-08-29
  • 基金资助:
    国家社科基金2025年度后期资助重点项目“以教育为方法:直面‘教育?人工智能’事件”(25FJKA002)。

How Education Responds to the Representational Crisis of Generative AI

Chengyu Yang   

  1. Institute of Curriculum and Instruction, East China Normal University, Shanghai 200062, China
  • Online:2026-09-01 Published:2026-08-29

摘要:

生成式人工智能的爆发标志着人类认知生态发生根本性的断裂,教育正经历从“检索范式”向“生成范式”的深刻转型。凭借概率计算与对自然语言的深度语义模拟,生成式人工智能在本体论层面上实质性地接管了表征的生产权。这种表征的深度外置化以及教育实践的算法中介化,将导致教育场域中意义理解、生命感知、情感联结与主体建构的整体失衡,进而撕裂出深层鸿沟,同时又隐含了超越人类中心主义的可能。面对上述“危机”,传统的表征主义与素朴的非表征理论均显现出解释力的贫乏。基于德勒兹的“差异性生成”哲学与德里达的“延异”理论,可构建出“生成性交织”的理论框架,并划定“算法的概率性重复”与“生命的差异性生成”边界。教育在对齐与逃逸的辩证法中,并非拒绝表征与人工智能,而是将机器生成的一般性统筹为教育的底层基底。作为一项延异技术,教育应以差异性叙事延迟意义判断,回应理解鸿沟;以感官锚点的重新校准延展身体判断,弥合生命鸿沟;以美育中的审美遭遇延迟情感定型,激活情动生成,跨越情感鸿沟;最终通过评价机制聚拢被算法分散的责任,重构游牧式主体,回应主体性鸿沟。

关键词: 生成式人工智能, 表征, 非表征, 表征外置化, 游牧式主体

Abstract:

The explosion of generative artificial intelligence (GenAI) marks a fundamental rupture in the human cognitive ecology, driving education through a profound transition from a “retrieval paradigm” to a “generative paradigm.” Relying on probabilistic computation and the deep semantic simulation of natural language, GenAI has substantially taken over the power of representational production at the ontological level. This deep externalization of representation, coupled with the algorithmic mediation of educational practices, will lead to a systemic imbalance in life perception, meaning comprehension, emotional connection, and subject construction within the educational field. Consequently, this tears open deep chasms while simultaneously implying the potential for de-anthropocentrism. Faced with this “crisis,” both traditional representationalism and naive non-representational theories reveal a poverty of explanatory power. Based on Gilles Deleuze’s philosophy of “differential becoming” and Jacques Derrida’s theory of “différance,” this paper constructs a theoretical framework of “generative intertwining,” demarcating the boundary between the “probabilistic repetition of algorithms” and the “differential becoming of life.” In the dialectic of alignment and flight, education does not reject representation and AI; rather, it integrates machine-generated generality as its underlying substrate. Furthermore, as a technology of différance, education should respond to the gap of understanding by using differential narratives to defer judgments of meaning; bridge the gap of life by recalibrating sensory anchors to extend bodily judgment; traverse the gap of emotion by activating affective generation through aesthetic encounters in aesthetic education; and ultimately respond to the gap of subjectivity by reforming evaluation mechanisms to regather the responsibility dispersed by algorithms and reconstruct the nomadic subject.

Key words: generative artificial intelligence, representation, non-representation, externalization of representation, nomadic subject