首页|生成式人工智能虚假信息协同共治研究

生成式人工智能虚假信息协同共治研究

扫码查看
原文链接
  • 万方数据
  • 维普
[研究目的]生成式人工智能虚假信息泛滥严重阻碍了互联网行业健康发展,协同共治有助于抑制生成式人工智能虚假信息.[研究方法]本研究从信息生成视角入手,构建"政府—生成式AI服务提供者—生成式AI服务使用者"三方协同共治演化博弈模型,进行仿真分析.[研究结论]研究结果表明,博弈主体协同共治策略受所有主体初始意愿影响;在成本支出方面,合理区间内降低技术治理成本和积极监管成本可促进生成式AI服务提供者和政府参与协同共治;在奖惩机制方面,政府奖励和惩罚措施可相互调节,且政府惩罚重心偏向生成式AI服务使用者时治理效果较好.同时,与政府惩罚相比,生成式AI服务使用者对提供者惩罚更加敏感,在合理区间内生成式AI服务提供者对政府奖励更为敏感;在社会形象方面,与公信力损失相比,政府对公信力提升更敏感.与声誉损失相比,生成式AI服务提供者对声誉提升更敏感.
Research on Collaborative Governance with Generative Artificial Intelligence in Disinformation
[Research purpose]Disinformation created by generative AI models seriously hinders the healthy development of the Internet industry,and collaborative governance helps each participating subject understand self-worth claims and effectively stops generative AI dis-information.[Research method]This study constructs a three-party collaborative governance evolution game model of"government-generative AI service provider-generative AI service user"from the perspective of information generation and conducts simulation analy-sis.[Research conclusion]For initial willingness,the collaborative strategies of game subjects are influenced by the initial willingness of all subjects.For cost,reducing the costs of technical governance and active regulation within a reasonable range can facilitate the engage-ment of generative AI service providers and governments in collaborative governance;In terms of rewards and penalties,government re-wards and penalties can regulate each other.It works better when the government penalizes the regulatory focus in favor of generative AI service users.At the same time,compared with government punishment,generative AI service users are more sensitive to the provider's penalties and within a reasonable range,generative AI service providers are more sensitive to government rewards than government punish-ment.For social image,governments are more sensitive to credibility gains than credibility losses.Generative AI service providers are more sensitive to reputation enhancement compared to reputation loss.

generative artificial intelligenceartificial intelligence generated contentdisinformationcollaborative governanceevolution-ary gamingstakeholders

郭海玲、卫金金、刘仲山

展开 >

河北大学管理学院 保定 071002

河北省数字治理与协同治理研究基地 保定 071002

生成式人工智能 人工智能生成内容 虚假信息 协同共治 演化博弈 利益相关者

国家社会科学基金青年项目河北省社会科学发展研究重点课题河北大学社会科学培育项目

17CTQ046202301030022022HPY012

2024

10.3969/j.issn.1002-1965.2024.09.015

情报杂志
陕西省科学技术信息研究所

情报杂志

CSTPCDCSSCICHSSCD北大核心
影响因子：1.502
ISSN：1002-1965
年,卷(期)：2024.43(9)
  • 22