| 孟玺,王玥,杨伟强,李庆霜,郭亚军.生成式AI环境下社交媒体信息迷雾的影响因素与生成路径——基于SEM与fsQCA的实证分析[J].数字图书馆论坛,2025,21(8):23~34 |
| 生成式AI环境下社交媒体信息迷雾的影响因素与生成路径——基于SEM与fsQCA的实证分析 |
| Influencing Factors and Generative Paths of Social Media Disinformation in the Context of Generative AI: Empirical Analysis Based on SEM and fsQCA |
| 投稿时间：2025-07-11 |
| DOI：10.3772/j.issn.1673-2286.2025.08.003 |
| 中文关键词: 生成式AI；社交媒体；信息迷雾；影响因素；SEM；fsQCA |
| 英文关键词: Generative AI; Social Media; Disinformation; Influencing Factor; SEM; fsQCA |
| 基金项目:本研究得到中央高校基本科研业务费基金项目“公安大数据智能分析过程中算法风险与规制研究”（编号：2024JKF21）、国家社会科学基金项目“算法分发环境下隐性数字弱势群体信息公平实现路径研究”（编号：22CTQ038）资助。 |
| 作者 | 单位 | | 孟玺 | 中国人民公安大学国家安全学院 | | 王玥 | 中国人民公安大学国家安全学院 | | 杨伟强 | 中国人民公安大学国家安全学院 | | 李庆霜 | 中国人民公安大学国家安全学院 | | 郭亚军 | 中国人民公安大学国家安全学院；郑州航空工业管理学院信息管理学院 |
|
| 摘要点击次数: 72 |
| 全文下载次数: 132 |
| 中文摘要: |
| 探究生成式AI环境下社交媒体信息迷雾的影响因素与生成路径，有助于完善我国网络空间安全治理体系，为网络信息生态优化提供理论参考。本研究基于信息生态理论、精细加工可能性模型与情绪感染理论构建社交媒体信息迷雾生成机制模型，采用结构方程模型（SEM）与模糊集定性比较分析（fsQCA）相结合的方法，对520份问卷数据进行实证分析。SEM结果显示，信息过载、交互氛围、感知个性化、信息有用性正向显著影响认知偏差和情绪感染，进而影响社交媒体信息迷雾生成，且认知偏差的路径系数高于情绪感染，是直接影响信息迷雾生成的最关键变量，感知拟人化通过情绪感染间接影响信息迷雾生成，信源可信度对社交媒体信息迷雾生成无显著影响；通过fsQCA方法分析发现3条触发社交媒体信息迷雾的路径。 |
| 英文摘要: |
| Exploring the influencing factors and generative paths of social media disinformation in the context of generative AI is helpful to improve China's cyberspace security governance system and provide theoretical references for network information ecology optimization. Based on the theory of information ecology, elaboration likelihood model, and emotional contagion theory, a model of social media disinformation generation mechanism is constructed. The structural equation model (SEM) and fuzzy set qualitative comparative analysis (fsQCA) are used to empirically analyze the data of 520 questionnaires. The SEM results show that information overload, interactive atmosphere, perceived personalization, and information usefulness positively and significantly affect cognitive bias and emotional infection, and then affect the generation of social media disinformation. The path coefficient of cognitive bias is higher than that of emotional infection, and it is the most critical variable that directly affects the generation of disinformation. Anthropomorphic perception indirectly affects the generation of disinformation through emotional infection, while the credibility of the information source has no significant impact on the generation of disinformation on social media. fsQCA method is used to analyze and find three paths that trigger the generation of social media disinformation. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
