【摘要】 以我国31个省级行政区的44份人工智能教育政策文本为研究对象，构建了适用于省域教育政策文本的量化分析路径，将LDA主题模型、社会语义网络分析与政策工具理论结合，量化分析政策的主题结构、工具特征与内在逻辑。研究提炼出政策顶层设计与区域实施等四大政策主题，发现政策工具呈现供给主导、环境支撑、需求薄弱的特征，文本以执行性“通知”为主。基于研究结果，从优化政策工具组合等多方面提出对策建议，为完善省级人工智能教育治理体系、推动区域教育数字化转型提供参考。更多还原

【Abstract】 Taking 44 artificial intelligence education policy texts from 31 provincial-level administrative regions in China as the research object,this paper constructs a quantitative analysis framework applicable to provincial education policy texts,by integrating the LDA topic model,social semantic network analysis and policy tool theory,conducts a quantitative analysis of the thematic structure,instrumental characteristics and internal logic of the policies. The study identifies four major policy themes including top-level policy design,regional implementation,etc.,and finds that the policy instruments feature a supply-dominant pattern with environmental support and insufficient demand-driven measures,and the policy texts are dominated by executive “notices”. Based on the research findings,countermeasures and suggestions are put forward from aspects such as optimizing the combination of policy instruments,providing a reference for improving the governance system of provincial artificial intelligence education and promoting the digital transformation of regional education.更多还原