人工智能治理范式变革——基于政策分析与主题建模的研究

Paradigm Shift in Artificial Intelligence Governance:A Study Based on Policy Analysis and Topic Modeling

人工智能治理范式是在人工智能技术治理中面对责任分配、资源配置、规则制定及挑战应对等问题时逐步形成的具有指导性、规范性及系统性的范例。文章通过引入范式变革理论,认为人工智能技术发展过程中不断产生的问题,以及人工智能治理导向、主体、内容的持续修正和调节需求是人工智能治理范式具有进步性和动态性的理论根源。基于此,对全球主要国家和国际组织发布的人工智能政策进行量化分析,通过量化政策发布机构、文件类型的时序变化,以及采用BERTopic主题建模归纳总结不同阶段人工智能治理内容的侧重点,解构人工智能治理范式的变革趋势,洞察人工智能治理复杂性。

The paradigm of artificial intelligence(AI)governance has emerged as a guiding,normative,and systematic framework to address challenges in responsibility allocation,resource distribution,rule-making,and problem-solving in AI governance.By introducing the theory of paradigm shifts,this study posits that the progressive and dynamic nature of AI governance paradigms is rooted in the continuous emergence of challenges in AI development and the ongoing need for adjustments in governance orientations,stakeholders,and content.This paper conducts a quantitative analysis of AI policies released by major countries and international organizations,examining the temporal evolution of policy-issuing institutions and document types.Using BERTopic modeling,it identifies and summarizes the shifting focal points of AI governance across different stages.By deconstructing the transformative trends of the AI governance paradigm,this study sheds light on the inherent complexity of AI governance.