Under the background of the development of artificial intelligence technology,the planning and dissemination of TV news should not only process a large amount of content quickly and deeply,extract structured information such as theme,label and style from information sources;but also learn to use audience behavior to explore and discover potential content,so as to realize personalized content presentation and recommendation.On the basis of the summary and analysis of the previous research results,the study expounds the methods and principles of semantic analysis and content recommendation;discusses the content production,interview and editing of TV news;proposes a research strategy of TV news communication mode.Through the empirical analysis,it is concluded that the AI-assisted TV communication mode can improve the dissemination effect of information,promote the optimization and integration of news resources,and enhance the audience's stickiness.