吉首大学学报(社会科学版) ›› 2026, Vol. 47 ›› Issue (4): 52-60.DOI: 10.13438/j.cnki.jdxb.2026.04.006

• 人工智能与大数据 • 上一篇    下一篇

生成式人工智能的认知卸载风险及其防范

齐绍平,郑君艳   

  1. (湘潭大学 马克思主义学院,湖南 湘潭 411105)
  • 出版日期:2026-07-01 发布日期:2026-07-16
  • 作者简介:齐绍平,男,博士,湘潭大学马克思主义学院教授,博士生导师;郑君艳,女,湘潭大学马克思主义学院博士生。

Cognitive Offloading Risks of Generative Artificial Intelligence and Their Mitigation

Qi Shaoping,Zheng Junyan   

  1. (School of Marxism,Xiangtan University,Xiangtan 411105,Hunan China)
  • Online:2026-07-01 Published:2026-07-16

摘要:生成式人工智能与人认知的深度融合正重塑现代认知生态,引发人机认知耦合的范式变革。当前,生成式人工智能主要引发三大认知卸载风险:一是思维惰性陷阱,即人工智能对信息处理与内容生成的替代,导致人主动思考与批判性思维退化;二是拟真欺诈悖论,即生成式模型凭借“逻辑自洽的虚构”混淆真实与虚假,瓦解社会信任的信息根基;三是认知操控困境,即算法通过信息筛选与排序等方式隐秘地塑造公众议程与价值判断,侵蚀人的决策自主性。针对上述风险,应当构建综合应对体系。要通过构建具身认知场域、强化深度思考训练,以重塑人思维的主动性与批判力,破除思维惰性;应依托监管立法协同、技术检测工具升级与个体信源核验机制,以重建真实可靠的信息生态,抵御拟真欺诈;要建立认知偏差反思机制、推动算法透明化交互设计,并引入“认知完整性”评估框架,以确保人在价值判断与意义建构中的主体地位,解构认知操控。总之,引导人机关系从工具性依赖走向智慧共生,在技术赋能与认知主权之间实现动态平衡,是数字文明时代维护认知生态健康、促进人类文明持续创新的关键路径。

关键词: 生成式人工智能, 认知卸载, 伦理嵌入, 技术异化, 过度依赖

Abstract: The deep integration of generative artificial intelligence with human cognition is reshaping the contemporary cognitive ecosystem,initiating a paradigm shift in human-AI cognitive coupling.Currently,generative AI primarily gives rise to three major cognitive offloading risks.The first is the trap of intellectual inertia:AI's substitution for information processing and content generation leads to the deterioration of active thinking and critical reasoning.The second is the paradox of realistic deception:generative models,by virtue of "logically self-consistent fabrications," blur the boundary between the real and the fabricated,thereby eroding the informational foundation of social trust.The third is the dilemma of cognitive manipulation:algorithms surreptitiously shape public agendas and value judgments through mechanisms such as information filtering and ranking,thereby undermining human decision-making autonomy.In response to these risks,a comprehensive countermeasure system should be established.To counteract intellectual inertia,it is necessary to cultivate embodied cognitive fields and strengthen deep-thinking training,thereby reshaping human initiative and critical capacity.To defend against realistic deception,a tripartite approach should be adopted:coordinated regulatory legislation,upgraded technological detection tools,and individual source-verification mechanisms,all aimed at reconstructing a reliable information ecosystem.To deconstruct cognitive manipulation,mechanisms for reflecting on cognitive bias should be developed,transparent algorithm-interaction designs should be promoted,and a "cognitive integrity" assessment framework should be introduced to ensure human agency in value judgment and meaning construction.Ultimately,guiding the human-AI relationship from instrumental dependence toward symbiotic wisdom,achieving a dynamic balance between technological empowerment and cognitive sovereignty,represents the crucial pathway for maintaining a healthy cognitive ecosystem and fostering sustained innovation in human civilization within the digital age.

Key words: generative artificial intelligence, cognitive offloading, ethical embedding, technological alienation, over-reliance

版权所有 © 2021《吉首大学学报(社会科学版)》编辑部
技术支持:北京玛格泰克科技发展有限公司
公众号 电子书橱 超星期刊 手机浏览 在线QQ