人工智能生成内容标识制度的逻辑更新与制度优化

Logical Updating and Institutional Optimisation of Artificial Intelligence Generated Content Labeling System

生成式人工智能正在颠覆人们获取信息的方式,但生成内容的质量良莠难辨。随着用户生成内容时代的分流治理机制失效,标识制度应成为人工智能内容治理的核心,发挥“生成来源+真伪判断”的双轨提示作用。占据信息优势地位的模型开发者、服务提供者应履行内容安全保障义务与信息质量披露义务,主动修正虚假信息引发的用户认知挤压、心理冲击等问题。在具体标识制度优化迭代中,应设计人工智能生成内容质量光谱标识,要求对生成来源、责任主体、内容质量进行强制标识,以激活标识制度在信息内容筛选的实质作用。

Generative artificial intelligence is disrupting people’s access to information,but the quality of generated content is hard to distinguish.With the failure of the triage governance mechanism in the era of user-generated content,the labeling system should become the core of AI content governance,and play the role of dual-track prompting of“generation source+authenticity judgement”.Model developers and service providers that occupy the information superiority position should fulfil the obligations of content safety guarantee and information quality disclosure,and take the initiative to correct the problems of user cognitive extrusion and psychological impact caused by false information.In the optimisation iteration of the specific marking system,the AI-generated content quality spectrum marking should be designed,requiring mandatory marking of the generation source,the responsible subject,and the content quality,in order to activate the substantive role of the marking system in the screening of information content.