The use of artificial intelligence (AI) to support academic work has recently been the subject of many debates. In this paper, we show that coupling an expert bibliometric system (ARTIREV) with a generative AI commonly used (ChatGPT) could help overcome the drawbacks of both systems to produce descriptive literature reviews. In a design research approach, and an experimental setting, we investigate the results obtained by each system alone, and the two combined, to support literature reviews on the same subject in two fields, Management and Medicine. Finally, we present open avenues of research into what we have termed the guided generative transformer (GGT) for science.
Read the publication
