网络舆情智能生成风险防控的PDRR体系构建  

The Construction of the PDRR System for Intelligent Generation and Risk Prevention of Online Public Opinion

作　　者：魏俊斌 Wei Junbin

机构地区：[1]哈尔滨工程大学马克思主义学院,哈尔滨150001

出　　处：《南京社会科学》2024年第6期76-87,共12页Nanjing Journal of Social Sciences

基　　金：国家社科基金青年项目“疫情背景下网络舆情的法治化治理研究”(20CKS045)的阶段性成果。

摘　　要：生成式人工智能在与用户的信息交互中重塑了网络舆情的生成与传播机制,诱发新型舆情风险,造成当下基于主体责任的人工智能治理范式无法有效应对网络舆情智能生成风险的责任主体多元化倾向,亟待构建一种全新的贯穿监管治理范式,实现治理价值的精准与动态安全转向。PDRR模型应用到网络舆情智能生成风险的治理能够有效弥补现有治理范式缺憾,为构建与生成式AI技术应用风险相适配的舆情信息全过程动态安全监管机制提供了可能,具有现实的基础和理论上的可行性。在PDRR模型安全政策指导下,通过将动态安全理念融入风险防控的顶层设计,以精准识别风险防控体系建立安全能力要素,并在实践层面建立以“贯穿监管”为重点的防护体系、以“应用安全”为核心的检测环节、以“责任联动”为抓手的应急处置、以“安全可控”为目标的系统恢复,构建一套覆盖网络舆情智能生成全生命周期的动态安全防护体系,有效防控生成式AI技术在网络舆情生成中的贯穿式风险,确保网络空间的信息安全。The generative artificial intelligence(AI)has reshaped the generation and dissemination mechanism of online public opinion through its interactions with users,and has induced new types of public opinion risks.This has rendered the current AI governance paradigm,based on subject responsibility,ineffective in addressing the increasingly diverse responsibility entities associated with the intelligent generation of online public opinion risks.It is urgent to construct a new regulatory governance paradigm that spans the entire process,which achieves precise and dynamic security transformation in governance values.The application of the PDRR model to the governance of intelligent generation risks of network public opinion can effectively make up for the deficiencies of the existing governance paradigms.It provides the possibility for the construction of a dynamic safety supervision mechanism for public opinion information throughout its lifecycle,adapting to the risks associated with the application of generative AI technology,and possesses both practical foundation and theoretical feasibility.Under the guidance of PDRR model safety policies,build a set of dynamic security protection system covering the whole life cycle of intelligent generation of online public opinion,by integrating dynamic security concept into the top-level design of risk prevention and control to accurately identify the security capability elements established by the risk prevention and control system,and at the practical level,establishing a protection system focusing on"through supervision",a detection link focusing on"application safety",an emergency responding with"responsibility linkage"as the starting point,and a system recovering with"safe and controllable"as the goal,which can effectively prevent and control the prevention risk of generative AI technology in the generation of online public opinion,and ensure the information security in cyberspace.

关 键 词：生成式人工智能 网络舆情 PDRR模型 防护体系 

分 类 号：G203[文化科学—传播学]