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

Ecological Security Risk and Governance of Network Information Content in AIGC Era——Take ChatGPT as an Example

[目的/意义]针对ChatGPT等人工智能生成内容(AI Generated Content,AIGC)兴起背景下,网络信息内容生成呈现出的新特点,梳理网络信息内容生态安全面临的新风险,分析新背景下进行网络信息内容生态安全治理的可行路径。[方法/过程]以总体国家安全观为指引,通过逻辑分析和以ChatGPT为对象的实验访谈案例分析,探析AIGC带来的网络信息内容生态安全风险。[结果/结论]AIGC是Web3.0阶段网络信息内容生成的主要方式,网络信息内容生成呈现出方式智能化、规模海量化、主体空心化以及形态多样化等新特点,而这一变化也会在资源、主体、基础设施以及外部辐射等四大风险域诱发12种典型风险。当前,可通过“动态监管—协同监管—韧性监管—多元监管—系统监管”相结合的“五位一体”监管思路,有效治理AIGC所带来的网络信息内容生态安全风险。

[Purpose/Significance]With the rise of artificial intelligent generated content(AIGC)such as ChatGPT,network information content generation has shown some new features.This paper summarizes the new risks faced by network information content ecological security,and analyzes the feasible paths of its governance under the new background.[Method/Process]Guided by the overall national security concept,this paper explored the ecological security risks of network information content brought by AIGC through logical analysis and the case analysis of experimental interview with ChatGPT as the object.[Result/Conclusion]AIGC is the main way of network information content generation in Web 3.0 stage.The generation presents new features such as intelligent mode,massive scale,hollow entities and diversified forms.This change also induces 12 typical risks in four risk domains of resources,entity,infrastructure and external radiation.The ecological security risks of network information content brought by AIGC can be effectively managed through the five-area integrated supervision idea of“dynamic supervision-collaborative supervision-resilient supervision-multiple supervision-systematic supervision”.