摘要
随着生成式人工智能在网络空间的广泛应用,网络信息内容生成呈现出自主化、规模化以及智能化等新特征。技术赋能内容生产的同时,也从内容供给、生成、传递等诸多环节,扰动原有的网络内容生态,由此引发的内容安全风险不断凸显。当前,采取集中式的风险治理模式难以及时回应技术变革带来的风险外溢,而行动者网络理论倡导的技术的社会形塑理念为生成式人工智能内容安全治理提供了新的思路。现阶段,应构建面向内容安全风险的可执行多重行动者网络体系,在具体治理路径上,围绕风险预防、识别、处置和重建,形成以风险预防为前置的制度性框架、以风险识别为核心的即时性治理机制、以风险消弭为目标的处置性责任链条以及以具体情境为支撑的内容重建机制,共同应对生成式人工智能带来的内容安全挑战。
With the widespread application of generative artificial intelligence in cyberspace,the generation of online information content has exhibited new characteristics such as autonomy,scale,and intelligence.While technology empowers content production,it also disrupts the original online content ecosystem from various aspects such as content supply,generation,and transmission,thereby highlighting content security risks.Currently,centralized risk governance models struggle to respond promptly to the spillover risks brought about by technological changes.In contrast,the actor network theory,which advocates for the social shaping of technology,offers new insights for the governance of content security in generative artificial intelligence.At this stage,it is essential to construct an executable multi-actor network system that addresses content security risks.In terms of specific governance paths,a system framework with risk prevention as a precondition,an immediate governance mechanism centered on risk identification,a responsibility chain aimed at risk elimination,and a content reconstruction mechanism supported by specific contexts should be formed to collectively address the content security challenges posed by generative artificial intelligence.
作者
赵依桐
李小波
ZHAO Yitong;LI Xiaobo(People's Public Security University of China,Beijing 100038,China)
出处
《河南警察学院学报》
2025年第1期102-116,共15页
Journal of Henan Police College
基金
2022年国家社会科学基金重点项目““算法政治推送”对我国意识形态安全的威胁与应对”(22AZZ011)。
关键词
生成式人工智能
内容安全风险
行动者网络
治理路径
generative artificial intelligence
content security risks
actor network
governance path
