我国人工智能政策语义、主题特征与演化研究——基于政策文本的量化分析

Purpose/significancePolicy as a special factor plays a guiding, regulating and restraining role in the development of China’s AI industry. This paper explores the current status of policy introduction and makes recommendations with a view to promoting the healthy development of AI in China.Method/processChina’s most representative 66 national and provincial AI policy texts are selected as research objects, and through BERT-Based hybrid LDA model modelling and semantic quantitative analysis methods, both their internal semantics and external attributes are comprehensively explored in terms of policy semantics, thematic features and their evolution.Result/conclusionAt present, China’s AI policy tools are mainly demand-based, and the policy theme is to promote domain integration, emphasizing basic technology research and actual results transformation. By analyzing the policy deficiencies, it is proposed that China needs to establish a coordinating mechanism for the overall planning of AI policies, explore new modes and paths for the transformation of practical benefits while paying attention to the governance of AI risks, and deepen the international cooperation and competition, among other policy recommendations.