This review assesses the performance of the AI tool “BoneView” in fracture detection, specifically examining its ability to improve diagnostic accuracy and reading efficiency for readers. It addresses the growing need for reliable support in radiographic interpretation amid rising fracture incidence and demand for radiologists through the use of AI. The research followed PRISMA guidelines and involved a systematic search of PubMed, Medline and Embase for studies comparing human fracture detection performance on radiographs with and without BoneView. Data on diagnostic accuracy and reading time were extracted and statistically analyzed, with quality assessed using QUADAS tools and results compared using McNemar and Z-tests. It was seen that the use of BoneView significantly improved fracture detection sensitivity in most cases, particularly among less experienced radiologists, across eight studies involving 98 readers, while specificity changes were mixed and generally less distinct. Diagnostic accuracy improved in several readers, especially those with intermediate or extensive experience which could be attributed to the fact that readers with experience tend to know what to look out for and are able to be more efficient with the use of BoneView. Region-specific analysis showed the largest gains in sensitivity for wrist, hand, and shoulder fractures. However, the methodological limitations and heterogeneity among studies highlight the need for further high-quality research.