图书情报工作

图书情报工作 >

2026 , Vol. 70 >Issue 10: 80 - 94

DOI: https://doi.org/10.13266/j.issn.0252-3116.2026.10.007

研究论文

AIGC技术对信息茧房效应的强化机制——基于情感偏向性的信息检索实证研究

  • 李鹏程 ,
  • 张继东 ,
  • 周文慧 ,
  • 李洁
展开
  • 1湖北工业大学经济与管理学院 武汉 430068;
    2湖北农村社会管理创新研究中心 武汉 430068;
    3武汉工商学院电子商务学院 武汉 430068;
    4山东大学图书馆 济南 250100
李鹏程，讲师，博士;张继东，教授，博士，博士生导师;周文慧，本科生;李洁，馆员，博士，通信作者，E-mail:allisonlijie@163.com。

收稿日期: 2025-04-08

  修回日期: 2025-08-02

  网络出版日期: 2026-05-07

基金资助

本文系国家社会科学基金青年项目“模糊认知视角下智慧图书馆资源推荐服务模式及实证研究”（项目编号：22CTQ009）和湖北工业大学高层次人才科研启动基金“层级语义功能感知下的学术文本相关工作章节自动生成研究”（项目编号：XJ2022009701）研究成果之一。

收起

Reinforcement Mechanism of Information Cocoon Effect by AIGC Technology - Empirical Study of Information Retrieval Based on Affective Biases

  • Li Pengcheng ,
  • Zhang Jidong ,
  • Zhou Wenhui ,
  • Li Jie
Expand
  • 1 School of Economics and Management, Hubei University of Technology, Wuhan 430068;
    2 Hubei Rural Social Management Innovation Research Center, Wuhan 430068;
    3 School of E-Commerce, Wuhan Business College, Wuhan 430068;
    4 Shandong University Library, Jinan 250100
Li Pengcheng, Lecture, PhD; Zhang Jidong, Professor, PhD, Doctoral Supervisor; Zhou Wenhui, Undergraduate Student; Li Jie, Librarian, PhD, Corresponding Author, E-mail: allison lijie@163.com.

Received date: 2025-04-08

  Revised date: 2025-08-02

  Online published: 2026-05-07

Supported by

This work is supported by the National Social Science Fund Youth Project “Empirical Research on Resource Recommendation Service Models in Smart Libraries from a Fuzzy Cognitive Perspective” (Grant No. 22CTQ009), and the Hubei University of Technology High-Level Talent Research Startup Fund Project “Research on the Automatic Generation of Relevant Chapters in Academic Texts Based on Hierarchical Semantic Function Perception” (Grant No. XJ2022009701).

Fold

摘要

[目的/意义]在AIGC技术重构信息生态系统的背景下，探究生成式人工智能对信息茧房效应的作用机制具有重要理论价值与社会意义。[方法/过程]以Google、ChatGPT-4o和Perplexity三种代表性信息检索工具为样本，设计并开展受控对照实验，构建和引入情感偏倚、情感极化、情感多样性等量化指标，多角度探究AIGC技术在增强检索结果同质化上的表现。在此基础上，结合案例分析，深入探讨AIGC的特征属性在调节信息流动和影响用户认知多样性上的作用机理。[结果/结论] 实验结果表明，相较于传统搜索引擎，AIGC类搜索平台更倾向于生成符合用户情感偏好的内容，其返回结果的情感一致性显著更强，表现为情感偏倚指数和极化指数显著提高，同时情感多样性显著降低。同时，AIGC通过其内部的“级联放大”机制在信息处理的各个环节中逐级放大情感偏向。此外，研究还识别出不同AIGC工具在实现该机制时存在“高偏倚输入”与“高强度过滤”的策略差异。

关键词： AIGC; 信息茧房效应; 信息检索; 情感多样性

本文引用格式

李鹏程 , 张继东 , 周文慧 , 李洁 . AIGC技术对信息茧房效应的强化机制——基于情感偏向性的信息检索实证研究[J]. 图书情报工作, 2026 , 70(10) : 80 -94 . DOI: 10.13266/j.issn.0252-3116.2026.10.007

Abstract

[Purpose/Significance] Against the backdrop of AIGC technology reconstructing the information ecosystem, exploring the mechanism of generative artificial intelligence on the information cocoon effect holds significant theoretical value and social implications. [Method/Process] This study selects three representative information retrieval tools - Google, ChatGPT-4o, and Perplexity - as samples. Through designing and conducting controlled comparative experiments, we construct and introduce quantitative indicators including emotional bias, emotional polarization, and emotional diversity to comprehensively investigate the performance of AIGC technology in enhancing the homogenization of search results. Furthermore, combining case analysis, we deeply examine the mechanism of AIGC’s characteristic attributes in regulating information flow and influencing users’ cognitive diversity. [Results/Conclusion] Experimental findings reveal that compared with traditional search engines, AIGC-based search platforms demonstrate stronger tendencies to generate content aligned with users’ emotional preferences. Their returned results exhibit significantly higher emotional consistency, manifested through notably increased emotional bias and polarization indices, alongside significantly reduced emotional diversity. Simultaneously, AIGC achieves this reinforcement effect through its internal “cascade amplification” mechanism, which amplifies emotional bias at each stage of information processing. In addition, the study identified differences in strategy between different AIGC tools in implementing this mechanism, namely “high bias input” and “high intensity filtering”.

Key words： AIGC; information cocoon effect; information retrieval; emotional diversity

参考文献

[1] 陆伟, 刘家伟, 马永强, 等. ChatGPT为代表的大模型对信息资源管理的影响[J]. 图书情报知识, 2023, 40(2): 6-9, 70. (LU W, LIU J W, MA Y Q, et al. The impact of large models represented by ChatGPT on information resources management[J]. Documentation, information & knowledge, 2023, 40(2): 6-9, 70.)
[2] 卢新元, 张进澳, 雷晓鹏. 人工智能生成内容环境下用户信息行为研究——以对话式搜索引擎为例[J]. 情报理论与实践, 2023, 46(12): 84-92. (LU X Y, ZHANG J A, LEI X P. A study on user information behavior in the context of AIGC: a case of conversational search engine[J]. Information studies: theory & application, 2023, 46(12): 84-92.)
[3] 刘智锋, 吴亚平, 王继民. 人工智能生成内容技术对知识生产与传播的影响[J]. 情报杂志, 2023, 42(7): 123-130. (LIU Z F, WU Y P, WANG J M. The impact of artificial intelligence generated content technology on knowledge production and dissemination[J]. Journal of intelligence, 2023, 42(7): 123-130.)
[4] 王丹, 姜艺, 陆伟. 数智时代信息服务中的隐私困境及应对策略——基于隐私让渡视角的探索[J]. 中国图书馆学报, 2024, 50(3): 67-81. (WANG D, JIANG Y, LU W. Privacy dilemmas and coping strategies in information services in the digital intelligence era: an exploration from the perspective of privacy waiver[J]. Journal of library science in China, 2024, 50(3): 67-81.)
[5] 王迁. 论人工智能生成的内容在著作权法中的定性[J]. 法律科学(西北政法大学学报), 2017, 35(5): 148-155. (WANG Q. On the legal characterization of AI-generated content in copyright law[J]. Legal science (Science of Law - Northwest University of Political & Law), 2017, 35(5): 148-155.)
[6] TONG R, YU W, ZHANG J, et al. Research on the application path of AIGC in higher education[C]//Wuhan international conference on e-business. Cham: Springer Nature, 2025: 259-269.
[7] WU S, LUO J, JIANG S, et al. A study on the credibility assessment of news content generated based on AIGC—using entertainment news as a case study[C]//20247th international conference on humanities education and social sciences. Amsterdam: Atlantis Press, 2024: 516-527.
[8] 张玥, 庄碧琛, 李青宇, 等. 同质化困境：信息茧房概念解析与理论框架构建[J]. 中国图书馆学报, 2023, 49(3): 107-122. (ZHANG Y, ZHUANG B C, LI Q Y, et al. The dilemma of homogenization: conceptual analysis and theoretical framework construction of the information cocoon[J]. Journal of library science in China, 2023, 49(3): 107-122.)
[9] 王静静, 叶鹰. 生成式AI及其GPT类技术应用对信息管理与传播的变革探析[J].中国图书馆学报, 2023, 49(6):41-50. (WANG J J, YE Y. Exploring the transformation of information management and communication by generative AI and its GPT-like technology applications[J]. Journal of library science in China, 2023, 49(6): 41-50.)
[10] MARTÍNEZ-ESPAÑA R, FERNÁNDEZ-PEDAUYE J, DE LUCÍA J G P, et al. Methodology for measuring individual affective polarization using sentiment analysis in social networks[J]. IEEE Access, 2024, 12: 102035-102049.
[11] 计卓琳, 吉久明. 网络用户情感茧房形成的影响因素研究[J]. 现代情报, 2024, 44(3): 70-80. (JI Z L, JI J M. Influencing factors on the formation of online users’ affective cocoons[J]. Journal of modern information, 2024, 44(3): 70-80.)
[12] 邱程程, 彭向斌, 胡城铭, 等. 信息茧房效应下用户情感极化影响因素及路径研究[J].图书馆学研究, 2024(7): 94-109. (QIU C C, PENG X B, HU C M, et al. Influencing factors and paths of user affective polarization under the information cocoon effect[J]. Research on library science, 2024(7): 94-109.)
[13] SUNSTEIN C R. Infotopia: how many minds produce knowledge[M]. New Work: Oxford University Press, 2006: 9.
[14] PARISER E. The filter bubble: what the Internet is hiding from you [M]. New York: The Penguin Press, 2011.
[15] HART W, ALBARRACÍN D, EAGLY A H, et al. Feeling validated versus being correct: a meta-analysis of selective exposure to information[J]. Psychological bulletin, 2009, 135(4): 555.
[16] 姜婷婷, 许艳闰. 国外过滤气泡研究: 基础, 脉络与展望[J]. 情报学报, 2021, 40(10): 1108-1117. (JIANG T T, XU Y R. Foreign research on filter bubbles: foundation, context and prospects[J]. Journal of the China Society for Scientific and Technical Information, 2021, 40(10): 1108-1117.)
[17] DUBOIS E, BLANK G. The echo chamber is overstated: the moderating effect of political interest and diverse media[J]. Information, communication & society, 2018, 21(5): 729-745.
[18] PIAO J, LIU J, ZHANG F, et al. Human–AI adaptive dynamics drives the emergence of information cocoons[J]. Nature machine intelligence, 2023, 5(11): 1214-1224.
[19] 徐翔. 公共信息茧房：社交媒体信息内容收敛现象与效应——基于新浪微博的文本挖掘[J]. 情报杂志, 2023, 42(1):113-123. (XU X. Public information cocoons: the convergence phenomenon and effects of social media content—a text mining analysis based on Sina Weibo[J]. Journal of intelligence, 2023, 42(1): 113-123.)
[20] 魏春丽, 徐建民, 张国防. 基于生成内容图解析的社交媒体用户信息茧房强度测度研究[J]. 情报杂志, 2025, 44(10): 172-180. (WEI C L, XU J M, ZHANG G F. Measuring the intensity of social media users’ information cocoons based on generated content graph analysis[J]. Journal of intelligence, 2025, 44(10): 172-180.)
[21] WANG L, KIM K. Analyzing group polarization through text emotion measurement and time series prediction: a comparative study across three online platforms[J]. Measurement: sensors, 2024, 33: 101216.
[22] CINELLI M, DE FRANCISCI MORALES G, GALEAZZI A, et al. The echo chamber effect on social media[J]. Proceedings of the national academy of sciences, 2021, 118(9): e2023301118.
[23] BERMAN R, KATONA Z. Curation algorithms and filter bubbles in social networks[J]. Marketing science, 2020, 39(2): 296-316.
[24] 张玥, 赵颖. 移动阅读环境下信息茧房成因探析[J]. 现代情报, 2021, 41(10): 3-11. (ZHANG Y, ZHAO Y. An exploration of the causes of information cocoons in the mobile reading environment[J]. Journal of modern information, 2021, 41(10): 3-11.)
[25] 傅诗婷, 姜婷婷, 田慧溢. 平衡信息接触能否帮助人们打破信息茧房? 信息感知与态度变化的关系研究[J]. 信息资源管理学报, 2023, 13(1): 41-51. (FU S T, JIANG T T, TIAN H Y. Can balanced information exposure help people break the information cocoon? A study on the relationship between information perception and attitude change[J]. Journal of information resources management, 2023, 13(1): 41-51.)
[26] 中国信息通信研究院. 人工智能生成内容(AIGC)白皮书(2022年)[R/OL]. [2024-03-23]. http://www.caict.ac.cn/kxyj/qwfb/bps/202209/P020220902534520798735.pdf. (China academy of information and communications technology. Artificial intelligence generated content (AIGC) white paper (2022) [R/OL]. [2024-03-23]. http://www.caict.ac.cn/kxyj/qwfb/bps/202209/P020220902534520798735.pdf.)
[27] WEI Y, TYSON G. Understanding the impact of ai-generated content on social media: The pixiv case[C]. Proceedings of the 32nd ACM international conference on multimedia. New Work: ACM, 2024: 6813-6822.
[28] 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, 68(18): 23-31. (DING Q J, SU J. Research on the optimization strategies for library information resource construction oriented towards generative artificial intelligence (AIGC)[J]. Library and information service, 2024, 68(18): 23-31.)
[29] DAVE T, AHTALURI S A, SINGH S. ChatGPT in medicine: an overview of its applications, advantages, limitations, future prospects, and ethical considerations[J]. Frontiers in artificial intelligence, 2023, 6: 1169595.
[30] 王正超. AIGC介入科研知识生产：风险表征、逻辑解构与责任重构[J].情报资料工作, 2025, 46(1): 57-67. (WANG Z C. AIGC’s intervention in scientific knowledge production: risk characterization, logical deconstruction, and responsibility reconstruction[J]. Information and documentation services, 2025, 46(1): 57-67.)
[31] 潘云鹤. 人和人工智能（AI）共同进化[J]. 研究与发展管理, 2023, 35(4): 1-4. (PAN Y H. Co-evolution of humans and artificial intelligence (AI)[J]. R&D management, 2023, 35(4): 1-4.)
[32] 梁怀新, 宋诚. AIGC时代的网络信息内容生态安全风险及其治理——兼以ChatGPT为对象的实验访谈案例分析[J]. 图书情报工作, 2023, 67(20): 58-69. (LIANG H X, SONG C. Ecological security risks and governance of network information content in the AIGC era: a case study of experimental interviews with ChatGPT[J]. Library and information service, 2023, 67(20): 58-69.)
[33] FANG X, CHE S, MAO M, et al. Bias of AI-generated content: an examination of news produced by large language models[J]. Scientific reports, 2024, 14(1): 5224.
[34] 郭顺利, 张雪宁, 苏新宁. 科研人员生成式人工智能应用信息茧房成因及实证研究[J].情报理论与实践, 2025, 48(2): 45-55. (GUO S L, ZHANG X N, SU X N. An empirical study on the causes of information cocoons for researchers using generative artificial intelligence[J]. Information studies: theory & application, 2025, 48(2): 45-55.)
[35] LI X Y, ZHANG Y F, MALTHOUSE E C. A preliminary study of ChatGPT on news recommendation: personalization, provider fairness, and fake news[C]//2023 international workshop on news recommendation and analytics. Singapore: CEUR-WS.org, 2023:3561.
[36] 曹树金, 曹茹烨. 从ChatGPT看生成式AI对情报学研究与实践的影响[J]. 现代情报, 2023, 43(4): 3-10. (CAO S J, CAO R Y. The impact of generative AI on information science research and practice from the perspective of ChatGPT[J]. Journal of modern information, 2023, 43(4): 3-10.)
[37] GRAVINA A G, PELLEGRINO R, PALLADINO G, et al. Charting new AI education in gastroenterology: cross-sectional evaluation of ChatGPT and perplexity AI in medical residency exam[J]. Digestive and liver disease, 2024, 56(8): 1304-1311.
[38] KEYLOCK C J. Simpson diversity and the Shannon–Wiener index as special cases of a generalized entropy[J]. Oikos, 2005, 109(1): 203-207.
Options
文章导航

/

〈 〉