A critical view on citation index a discussion with ChatGPT

Abstract

This “opinion article” presents the “dialogue”, namely the series of questions addressed to chatGPT and its outputs, on the question of citation index and other metrics associated with scientific publication. The AI is asked to explain basic quantities (citation index, impact factor of journals, h-index, etc) and through different questions, to give outputs on different important problems, mainly related with citation index and its biases. The AI clearly expresses that the citation index alone should not be used for evaluation purposes, and is led to predict that in certain situations a high quality scientific article would be with a low citation index. The AI also replies that the citation index depends on several factors in a complex way, and is probably not defined in an unambiguous way from a mathematical point of view, but it appears unable to explain evolutions in the field of informetrics that occurred following issues with citation index. The article concludes that the citation index is irrelevant as it is not normalized and strongly depends on the publishing journal and databases, and that it is potentially dangerous for science.