Retained surgical items (RSIs) remain a persistent patient safety problem. This review examined the use of machine learning (ML) and deep learning (DL) tools to support RSI prevention and detection and how these applications can be integrated with existing safety practices in perioperative care.