摘要
[研究目的]随着生成式人工智能(GAI)技术在多模态领域的快速迭代,加剧了信息内容生成和传播的风险。分析GAI技术引发的网络舆情风险特征,实现有效风险预警,是未来舆情治理中亟待解决的重要工作。[研究方法]聚焦于由GAI引发的网络舆情事件,基于引爆点理论,从网络舆情内容、参与者和环境三个维度构建风险评价指标体系,通过熵权法和VIKOR法确定指标权重并对风险进行综合评价,同时结合舆情出现的时间节点分析出风险发生的关键节点,最后对分析结果进行了验证,实现了对舆情风险的有效预警。[研究结果/结论]研究发现,内容维度是舆情风险的主要影响因素,舆情风险发生的关键节点处于2024年1月至3月间。通过指标构建、综合评价构建风险预警模型,实现了对舆情风险的量化研究,能够有效识别由GAI引发的网络舆情风险,为未来预警该类风险提供理论支持。
[Research purpose]With the rapid iteration of GAI technology in the field of multimodality,the risk of information content generation and dissemination has intensified.Analyzing the characteristics of public opinion risks triggered by GAI technology and achieving effective risk early warning is an important task that urgently needs to be addressed in future public opinion management.[Research method]This paper focuses on public opinion events triggered by GAI.Based on the tipping point theory,it constructs a risk evaluation index system from three dimensions:content of public opinion,participants,and environment.Using the entropy weight method and VIKOR method,it determines the weights of the indicators and conducts a comprehensive risk evaluation.Additionally,it analyzes the critical nodes of risk occurrence by examining the time nodes of public opinion emergence.Finally,it validates the analysis results,achieving effective early warning of public opinion risks.[Research result/conclusion]This paper finds that the content dimension is the main influencing factor of public opinion risk,with the critical nodes of risk occurred from January to March 2024.This paper achieves a quantitative study of public opinion risks by constructing indicators and conducting a comprehensive evaluation to build a risk early warning model.This model effectively identifies public opinion risks triggered by GAI,providing theoretical support for future early warning of such risks.
作者
黄炜
吴昀璟
余辉
张瑞
Huang Wei;Wu Yunjing;Yu Hui;Zhang Rui(School of Economics and Management,Hubei University of Technology,Wuhan 430068;Business School of Wuhan Qingchuan University,Wuhan 430204;Hubei Digital Industrial Economy Development Research Center,Wuhan 430068)
出处
《情报杂志》
北大核心
2025年第3期98-106,共9页
Journal of Intelligence
基金
湖北省高等学校哲学社会科学研究重大项目“新时代高校突发事件网络舆情分析与引导机制研究”(编号:19ZD025)
湖北省科技创新人才计划项目“基于北斗位置服务的危化企业安全智造平台”(编号:2023DJC056)研究成果。
