Thematic Analysis and Evolution Process of US Defense Artificial Intelligence Strategy
In the face of a series of AI strategic initiatives initiated by the US Department of Defense,identifying the themes and evolutionary characteristics of the US defense AI strategy based on the perspective of text mining can provide a more objective understanding of the substantive intent of the US defense AI strategy and provide a reference for China's defense AI strategic planning.Firstly,the theme of US defense AI strategy(2018-2022)through high-frequency by unstructured Text mining tools.Secondly,the evolution characteristics of the US defense AI strategy were observed,by drawing a co-occurrence network graph of the theme words in the US defense artificial intelligence strategy text.Finally,suggestions are proposed for Chinese national defense AI strategic planning based on this.The development of artificial intelligence strategy by the US Department of Defense can be divided into five major themes:strategic entities,technological innovation,ethical security,strategic resources,and diversified cooperation.In the evolution process,the US national defense AI system was built by JAIC as the core sector with multiple departments,from terms of strategic entities.In terms of technological innovation,it has evolved from self-built artificial intelligence combat systems to establishing an integrated land,sea and air command and control network.Regarding strategic resources,data is a key asset,talent is a key element,and funding is a key assistance.In terms of ethical security,behind the transition from"AI ethical standards"to"responsible AI"is the transition from"reflecting democratic values"to"clearing public opinion barriers for US military hegemony".In terms of diversified cooperation,domestically,the public and private sectors built the AI technology ecosystem in the United States.Externally,the global RAI ecosystem was built by the US and allied partner countries.
US Department of DefenseArtificial Intelligence StrategyText MiningThematic AnalysisEvolutionary Characteristics
万方数据
维普
