Artificial intelligence in veterinary diagnostic histopathology: Concepts, applications, and clinical implementation

Artificial intelligence (AI) in human diagnostic histopathology has advanced rapidly, but clinical adoption has lagged behind research. Veterinary diagnostic pathology, being less regulated, could adopt AI faster, but the extent of its integration into diagnostic laboratories is unknown. This article combines a narrative introduction to key concepts, a scoping review of published applications, and an exploratory survey of diagnostic laboratories to provide an overview of AI in veterinary diagnostic histopathology. The narrative review highlights emerging approaches, including vision transformer-based foundation models and self-supervised learning, which may improve generalizability and reduce data requirements, although adaptation to the veterinary context remains challenging. The scoping review identified 34 studies published through a September 5, 2025, literature search, indicating steady but limited growth. Most used convolutional neural networks for relatively low-level tasks such as cell quantification, area measurement, and diagnosis. The survey had a low response rate (2 of 5 laboratories). Both respondents recognized potential benefits of AI. One laboratory had incorporated 2 AI models into routine diagnostic workflows and was evaluating additional applications, while the other had not yet implemented AI but was actively considering several models. Reported barriers included generalizability, latency, and cost. Overall, AI research in veterinary diagnostic histopathology is growing, but clinical adoption remains incompletely characterized. While less restrictive regulation may allow faster uptake than in human medicine, the lack of oversight places responsibility for validation on laboratories and pathologists. For trustworthy implementation, veterinary pathologists must understand AI fundamentals, and the field should prioritize standards for validation, transparency, and reporting.