Of great practical importance is the problem of analyzing existing and creating new educational materials: texts, drawings, mental maps, diagrams, presentations, animations, various tasks, etc. In general, any teaching materials on physics and astron-omy can be reduced to a text containing a certain number of relevant concepts, princi-ples, laws, educational tasks, questions, etc. The article discusses the problem of analyz-ing educational texts and identifying their key characteristics, which is important for creating high-quality teaching materials and improving teaching methods. The purpose of the work is to explore the possibilities of online resources for analyzing educational texts, determining their parameters and creating new educational materials. The article considers online platforms that allow to automatically solve various problems automati-cally: from identifying the main concepts and thoughts in educational texts, establishing inter- and intra-subject connections to building semantic networks that show the con-nections between key concepts. Practical aspects of the use of neural network technolo-gies are discussed. The possibilities of artificial intelligence in creating educational ma-terials are analyzed, including optimal structuring of educational information, descrip-tion of experimental evidence of theories, solving qualitative and quantitative problems, obtaining a list of scientific concepts associated with this concept. Examples of queries to a neural network are presented. The article considers Internet resources that help to find the average length of words and sentences, evaluate the Flash readability index and Fog index, and calculate the structural complexity of a text.
R. Mayer (Fri,) studied this question.