Disease and Health Surveillance in Companion Animals Using Artificial Intelligence and Machine Learning

Companion animal disease surveillance now benefits from collated databases of electronic health records and artificial intelligence. This review examines computational approaches for analyzing unstructured veterinary clinical text, from rule-based systems through traditional neural networks to modern transformer models. Domain-adapted encoders like PetBERT enable efficient disease coding and syndromic surveillance, while generative models offer new capabilities. Topic modeling provides unsupervised pattern discovery. Key challenges include model generalization across clinical settings, privacy protection through deidentification, standardized evaluation frameworks, and environmental sustainability. Strategic deployment of appropriately sized models can advance One Health surveillance while respecting environmental responsibility.