The use of artificial intelligence in the implementation of circular economy practices in reverse logistics – a systematic mapping study

Purpose In an era of deteriorating environmental conditions, increasing attention is being paid to incorporating circular economy principles and reverse logistics into the management of End-of-Life and End-of-Use products. This article presents the results of a literature re-view on the potential use of artificial intelligence, deep learning, and machine learning to implement the 10R practices of the circular economy in reverse logistics. Design/methodology/approach The objective of the article was achieved through a systematic mapping study, which allowed for the identification and categorization of selected articles in terms of reverse logistics, the most commonly used circular economy practices, and the scope of application of artificial intelligence, deep learning, and machine learning in these areas. Findings The findings of the study indicate that the topic has received limited scholarly attention to date. The use of circular economy practices in reverse logistics focuses on recycling, remanufacturing, and reuse processes, while the use of artificial intelligence, deep learning, and machine learning concerns operational issues such as goods identification and general aspects of artificial intelligence implementation in reverse logistics, e.g. barriers. Research limitations/implications The research results presented certainly have their limitations. The first of these concerns the focus on a single database, Scopus. Expanding the scope of the search to include other databases such as Web of Science, ProQuest, or IEEE Xplore would increase the scope of the analyzed material. The second limitation concerns the phrases used to identify articles, which in the field of reverse logistics could be expanded to include circular supply chains, closed-loop supply chains, or green supply chains. Furthermore, in terms of neural network algorithms, the search for articles that mention AI, ML, and DL could be expanded to include diverse types of neural networks. The CE practices themselves could be expanded to include additional ones, based, for example, on the 21R model (Tsironis et al., 2024). Another limitation concerns the restriction of the searched articles to those written in English only. It should also be noted that the Mendeley application, which was used to analyze the metadata, does not have the full capabilities to examine them, which certainly limits the cognitive process. Practical implications The presented research results can support operational management areas such as waste management or green supply chain management. They should also draw attention to the intensification of the implementation of circular economy practices and reverse logistics using artificial intelligence in organizational management. They should also prompt reflection on changing existing business models (often based on a linear economy). Social implications The topic addressed is very important from a social point of view. Environmental degradation caused by industries that have been using linear economy principles for decades is leading to a climate catastrophe. Addressing this issue may slow down this process and lead to more intensive work on solutions that will not only offset the negative impact of industry on the environment, but also actively support it by contributing to the achievement of sustainable development goals. This will improve the quality of life for entire communities. Originality/value The research clearly indicates there is a research gap in this area. By expanding our cur-rent understanding of the potential applications of artificial intelligence, deep learning, and machine learning in implementing circular economy practices in reverse logistics, these findings provide a solid starting point for designing and conducting research in this field.