This systematic literature review explores the intersection of Artificial Intelligence (AI) and Predictive Maintenance (PdM) within Industry 4.0. Using a PRISMA-based methodology, 123 studies published between 2014 and April 2024 were analyzed to characterize technological trends, algorithmic choices, industrial applications, and evaluation practices. The review reveals a consistent growth of research interest, driven by the widespread adoption of Internet of Things (IoT) devices and increased data availability. The manufacturing sector dominates the literature, although most studies rely on standardized datasets rather than real industrial environments. Among the identified AI methods, Random Forest (RF), Support Vector Machine (SVM), Decision Tree (DT) and K-Nearest Neighbors (KNNs) represent the most frequently applied algorithms for tasks such as failure prediction, fault detection, and remaining useful life (RUL) estimation. Model performance is commonly evaluated with Accuracy (Acc), Precision, Recall, F1-Score, and Root Mean Square Error (RMSE), reflecting the prevalence of both classification and regression-based PdM analyses. Despite significant advances, this review identifies persistent gaps, including limited domain diversity, scarce long-term real-world validation, and insufficient use of eXplainable AI (XAI) techniques. The findings highlight the need for broader domain coverage, improved interpretability, and validation under realistic industrial conditions. Overall, this review consolidates current knowledge on AI-enabled PdM and outlines critical directions to enhance reliability, transparency, and industrial relevance in the context of Industry 4.0.
Arez et al. (2026) studied this question.