One of the new educational technologies that has shown its undoubted effectiveness is e-learning. In developed countries, e-learning covers all levels of education and is widely used not only in universities, but also in high school and in the organization of corporate (postgraduate) education [1]. Such platforms require the elaboration of high-quality and relevant teaching resources, the constant updating of existing ones. This, in turn, is a complex process consisting of processing a variety of materials, their analysis, synthesis, creative development and processing of all elements to build a single harmonious structure [2]. Up to now, far too little attention has been paid to dynamic content generation for e-learning courses. In our previous work [3], we have proposed a program model for the dynamic creation of training courses and discussed the ways of content extraction and character recognition from images such as PDF, DOC, DOCX, and HTML. This paper continues our research and regards extracted text as meaningful pieces of information. Further on, using the NLP approach, we build some models to train our machine to make associations between a particular input and its corresponding output. An AI system uses statistical analysis methods to build its own “knowledge bank” and discern which features best represent the texts before making predictions for unseen data. A constructed system of AI, by applying semantic and pragmatic analysis, selects the most common parts of information and builds some new coherent text. This text is given in the text editor for further editing and downloading. This article was written within the framework of the research project 20.80009.5007.22 Intelligent information systems for solving ill-structured problems, processing knowledge and big data. Link to project website: http://www.math.md/en/projects/20.80009.5007.22/.