The author reviews the scientific and practical conference on artificial intelligence (AI) in libraries held by RAS Institute for Scientific Information for Social Sciences. He summarizes the presented papers exploring theoretical and practical aspects of using AI tools in library services in Russia and neighbor states. In particular, the author focuses on using artificial neural networks for information retrieval, e-library concept processing (classification, abstracting and annotating), full text recognition, uniform cataloguing, textbook conversion into online courses, and design of reading lists systems. At the conference, the speakers argued that the neural networks were slowly implemented in Russian libraries due to the lack of tagged “library” data sets in open access. They also emphasized the role of libraries in teaching users to work with AI applications. The conference participants consented on the fact of rapid development of this vector in the libraries. The librarians should be more persistent in using available universal applications, partner with Russian large IT-companies, and undertake to design specialized AI-tools for deep content processing of digitized information arrays.