图书情报工作

图书情报工作 >

2026 , Vol. 70 >Issue 1: 94 - 103

DOI: https://doi.org/10.13266/j.issn.0252-3116.2026.01.008

研究论文

生成式人工智能中“信息茧房”的生成与应对

  • 黄立赫
展开
  • 河南警察学院 马克思主义学院 郑州 450046
黄立赫，讲师，博士，E-mail：370299367@qq.com。

收稿日期: 2024-11-25

  修回日期: 2025-03-12

  网络出版日期: 2026-01-16

基金资助

本文系2024年度河南省高等教育教学改革研究与实践项目“工程训练实践课程思政建设质量评价和激励机制的探索与研究”(项目编号:2024SJGLX0332)和2023年河南省高等教育教学改革研究与实践项目(研究生教育类)“工程类专业学位研究生实践能力与职业素养提升策略研究与实践”(项目编号:2023SJGLX112Y)研究成果之一。

收起

The Formation and Mitigation of Information Cocoons in Generative Artificial Intelligence

  • Huang Lihe
Expand
  • School of Maxism, Henan Police College, Zhengzhou 450046

Received date: 2024-11-25

  Revised date: 2025-03-12

  Online published: 2026-01-16

Supported by

This work is supported by the 2024 Higher Education Teaching Reform Research and Practice Project in Henan Province,titled “Exploration and Research on the Quality Evaluation and Incentive Mechanism of Ideological and Political Education in Engineering Training Practice Courses”(Grant No. 2024SJGLX0332), and the 2023 Higher Education Teaching Reform Research and Practice Project in Henan Province (Graduate Education Category), titled “Research and Practice on Strategies for Enhancing the Practical Abilities and Professional Competencies of Engineering Professional Degree Graduate Students”(Grant No. 2023SJGLX112Y).

Fold

摘要

[目的/意义] Chat GPT等生成式人工智能工具具有交互性、主动性和智能性，可以显著优化信息获取效率并提高内容精准度，但同时隐含的信息茧房现象却容易被忽视，因此需要探讨生成式人工智能中信息茧房的形成机制、表现形态及其潜在风险并提出应对策略。[方法/过程] 利用定性研究方法，基于反馈环路理论等，分析信息茧房的生成逻辑和机制，识别其主要形态，分析其主要风险，提出破解策略。[结果/结论] 信息茧房共有机器认知共振、反馈环路加强、个体选择偏向三种形态。其生成逻辑是通过持续“学习”和“定制”用户的信息偏好，加速用户在信息消费过程中的顺从倾向，从而形成“信息茧房”，这将带来多种问题，包括但不限于认知偏误、懒人思维和公共讨论空间压缩、甚至引发人类思维退化等风险。为了规避这一风险的产生，需要从培育个体批判性思维等角度进行引导，实现人与机器的“双向奔赴”。

关键词： 生成式人工智能; ChatGPT; 信息茧房

本文引用格式

黄立赫 . 生成式人工智能中“信息茧房”的生成与应对[J]. 图书情报工作, 2026 , 70(1) : 94 -103 . DOI: 10.13266/j.issn.0252-3116.2026.01.008

Abstract

[Purpose/Significance] Generative artificial intelligence tools such as ChatGPT, characterized by interactivity, proactivity, and intelligence, significantly enhance information retrieval efficiency and content precision. However, their implicit role in fostering “information cocoons” remains understudied. This research aims to investigate the formation mechanisms, manifestations, and potential risks of information cocoons in generative AI systems, while proposing mitigation strategies. [Method/Process] Based on feedback loop theory, this study adopted qualitative research methods to analyze the generative logic and mechanisms of information cocoons, identified their primary forms, evaluated associated risks, and developed targeted countermeasures. [Result/Conclusion] In human-AI interactions, information cocoons present three main forms: machine-cognitive resonance, reinforcement of feedback loops, and individual selection bias. These phenomena stem from generative AI’s “unidirectional compliance” design. By continuously “learning” and “customizing” users’ information preferences, generative AI accelerates users’ tendency toward compliance in the information consumption process, which in turn deepens the formation of the “information cocoon.” Risks include cognitive biases, intellectual complacency, erosion of public discourse, and long-term degradation of human critical thinking. To mitigate these risks, strategies emphasize cultivating individual critical thinking and fostering two-way collaboration between humans and machines, balancing AI’s efficiency with human oversight to achieve symbiotic innovation.

Key words： generative artificial intelligence; ChatGPT; information cocoon

参考文献

[1] 中共中央关于进一步全面深化改革、推进中国式现代化的决定[N].人民日报, 2024-07-22(1).
[2] 丁遒劲,苏静.面向生成式人工智能(AIGC)的图书馆信息资源建设优化策略研究[J].图书情报工作, 2024(18):23-31.(DING Q J, SU J. Research on optimization strategies of library information resource construction oriented to generative artificial intelligence (AIGC)[J]. Library and information service,2024(18):23-31.)
[3] 张玲玲,黄务兰.基于ChatGPT API和提示词工程的专利知识图谱构建[J].情报杂志, 2025(3):180-187.(ZHANG L L,HUANG W L. Construction of patent knowledge graph based on ChatGPT API and prompt engineering[J]. Journal of intelligence,2025(3):180-187.)
[4] 杨俊,谭丰隆,陈婧.从ChatGPT到"LibGPT ":生成式人工智能驱动的新一代图书馆[J].图书情报工作, 2024(9):3-12.(YANG J, TAN F L, CHEN J. From ChatGPT to" LibGPT":the new generation library driven by generative artificial intelligence[J]. Library and information service, 2024(9):3-12.)
[5] VERMA M. Novel study on AI-based chatbot (ChatGPT) impacts on the traditional library management[J]. International journal of trend in scientific research and development (IJTSRD),2023, 7(1):961-964.)
[6] 王宗水,谷迎晓,赵红.基于Chat GPT的内容生成与知识传播模式创新[J].图书馆, 2023(6):42-48.(WANG Z S, GU Y X,ZHAO H. Content generation and knowledge dissemination model innovation based on ChatGPT[J]. Library, 2023(6):42-48.)
[7] 王静静,叶鹰,王婉茹. GPT类技术应用开启智能信息处理之颠覆性变革[J].图书馆杂志, 2023(5):9-13.(WANG J J, YE Y, WANG W R. GPT technology application initiates disruptive reform in intelligent information processing[J]. Library journal,2023(5):9-13.)
[8] PANDA S, KAUR N. Exploring the viability of ChatGPT as an alternative to traditional chatbot systems in library and information centers[J]. Library hi tech news, 2023, 40(3):22-25.
[9] 周旭.机遇与挑战:Chat GPT普及背景下图书馆的应对分析[J].图书馆, 2023(6):34-41, 48.(ZHOU X. Opportunities and challenges:analysis of library responses in the context of ChatGPT popularization[J]. Library, 2023, 345(6):34-41,48.)
[10] 李旭光,胡奕,王曼,等.人工智能生成内容研究综述:应用、风险与治理[J].图书情报工作, 2024(17):136-149.(LI X G, HU Y, WANG M, et al. A review of research on AI-generated content:applications, risks, and governance[J]. Library and information service, 2024, 68(17):136-149.)
[11] GUESS A, NYHAN B, LYONS B, et al. Avoiding the echo chamber about echo chambers[J]. Knight foundation, 2018,2(1):1-25.
[12] BISWAS S S. Role of ChatGPT in public health[J]. Annals of biomedical engineering, 2023, 51(5):868-869.
[13] KING M R, ChatGPT. A conversation on artificial intelligence,chatbots, and plagiarism in higher education[J]. Cellular and molecular bioengineering, 2023, 16(1):1-2.
[14] WOLLEBÆK D, KARLSEN R, STEEN-JOHNSEN K,et al. Anger, fear, and echo chambers:the emotional basis for online behavior[J]. Social media society, 2019, 5(2):2056305119829859.
[15] LIEBRENZ M, SCHLEIFER R, BUADZE A, et al. Generating scholarly content with ChatGPT:ethical challenges for medical publishing[J]. The Lancet digital health, 2023, 5(3):e105-e106.
[16] THORP H H. ChatGPT is fun, but not an author[J]. Science,2023, 379(6630):313.
[17] PATEL S B, LAM K. ChatGPT:the future of discharge summaries?[J]. The lancet digital Health, 2023, 5(3):e107-e108.
[18] VAN DIS E A M, BOLLEN J, ZUIDEMA W, et al. ChatGPT:five priorities for research[J]. Nature, 2023, 614(7947):224-226.
[19] YUSUF N, AL-BANAWI N, AL-IMAM H A R. The social media as echo chamber:the digital impact[J]. Journal of business&economics research, 2014, 12(1):1-10.
[20] NGUYEN C T. Echo chambers and epistemic bubbles[J].Episteme, 2020, 17(2):141-161.
[21] CHOI D, CHUN S, OH H, et al. Rumor propagation is amplified by echo chambers in social media[J]. Scientific reports, 2020,10(1):310.
[22] DIAZ-DIAZ F, SAN MIGUEL M, MELONI S. Echo chambers and information transmission biases in homophilic and heterophilic networks[J]. Scientific reports, 2022, 12(1):9350.
[23] HABERMAS J. The structural transformation of the public sphere:an inquiry into a category of bourgeois society[M].Cambridge:MIT Press, 1989.
[24] BOHMAN J, REHG W (Eds.). Deliberative Democracy:Essays on Reason and Politics[M]. Cambridge:MIT Press, 1997.
[25] LIVINGSTONE S. Engaging with media--a matter of literacy?[J].Communication, culture&critique, 2008, 1(1):51-62.
[26] PAFFENHOLZ T, SPURK C. Civil society, Civic engagement,and peacebuilding[R]. Social Development Papers:Conflict Prevention and Reconstruction, Washington DC:World Bank,2006:10-55.
[27] POWELL III A C. Falling for the Gap:Whatever happened to the digital divide[C]//COMPAINE B M. The digital divide:facing a crisis or creating a myth. Cambridge:MIT Press, 2001:309-314.
[28] LANG A, CHUNG Y, LEE S, et al. It's an arousing, fast-paced kind of world:the effects of age and sensation seeking on the information processing of substance-abuse PSAs[J]. Media psychology, 2005, 7(4):421-454.
[29] KELLNER D, SHARE J. Toward critical media literacy:core concepts, debates, organizations, and policy[J]. Discourse:studies in the cultural politics of education, 2005, 26(3):369-386.
[30] 黄洁梅."对抗性"减弱:生成式人工智能阅读交互与人类思维弱化[J].情报探索, 2023(12):44-50.(HUANG J M. Weakening of "Antagonism": generative artificial intelligence reading interaction and the weakening of human thinking[J]. Information research, 2023(12):44-50.)
[31] MEZIROW J. Transformative learning theory[C]//TAYLOR E W.Transformative learning in practice:insights from community,workplace, and higher education. San Francisco:Jossey-Bass,2009:18-31.
[32] HEIDEGGER M. The question concerning technology and other essays[M]. LOVITT W, trans. New York:Harper&Row, 1977:23-45.
[33] POPPER K R. The logic of scientific discovery[M]. London:Hutchinson&Co., 1959:65-76.
[34] WELSBY P, CHEUNG B M Y. ChatGPT[J]. Postgraduate medical journal, 2023, 99(1176):1047-1048.
[35] LIU Z, ZHANG Y, WANG X, et al. Human--AI adaptive dynamics drives the emergence of information cocoons[J].Nature machine intelligence, 2023, 5(7):521-533.
[36] YIN J. Interactive subjectivity of generative artificial intelligence:A perspective from autopoietic systems theory[J]. Social sciences in China, 2024, 45(3):89-107.
[37] ZHAO X, JIANG Z. Endogenous security governance of generative AI:a new paradigm for breaking information cocoons[J]. Journal of communication, 2023, 73(6):102-118.
[38] LI H, CHEN W. Algorithmic domestication and cognitive lock-in effects:evidence from short video platform users[J]. New media&society, 2024, 26(1):345-367.
Options
文章导航

/

〈 〉