Semantic Document Indexing with Generative AI

The possibilities of processing abstract information using generative artificial intelligence systems, in particular ChatGPT, are presented to solve the problems of generating semantic maps, semantic indexing, analysis and visualization, which makes it possible to consider such systems as a useful analytical tools. The ChatGPT system was used to automatically extract basic concepts from documents from a thematic array on the topic of information technology in construction, i.e. perform semantic indexing, as well as build a semantic network for the selected subject area. A working model has been implemented that allows you to find relevant records by clicking on nodes and edges of the constructed semantic network, i.e. navigate through the source information array.