From START to Perplexity: the evolution of artificial intelligence systems in the information and library sphere

Abstract

The evolution of the application of artificial intelligence systems in information and library activities is considered, starting with the first semantic search system START developed in 1993 and ending with large generative language models such as ChatGPT and Perplexity in the 2024 versions. A detailed description of all varieties is given, indicating the developments that have become milestones in the evolution of intelligent systems. The level of application of neural networks in reference and bibliographic work is analyzed, including the preparation of analytical reviews, digitization of retrospective publications, individual assistants for recommended reading, collection verification, home delivery of necessary sources to users, and robotic storage systems. It is noted that the process of cataloguing using artificial intelligence is possible only for documents presented in digital form and is most beneficial when compiling descriptions of similar publications.