Imagine a world where AI can instantly diagnose illnesses from complex veterinary records, streamlining diagnoses and potentially saving lives. That's the promise of groundbreaking new research from Colorado State University (CSU). Traditionally, translating messy, free-text veterinary notes into standardized medical codes (SNOMED-CT) has been a laborious, manual task. This bottleneck hinders crucial research and the integration of animal health data into broader "One Health" initiatives, which recognize the interconnectedness of human, animal, and environmental health. The CSU team tackled this challenge by leveraging the power of large language models (LLMs), the same technology behind AI chatbots. They fine-tuned several freely available LLMs, including GatorTron (trained on a vast dataset of human medical records) and VetBERT (trained on veterinary records), to decode free-text veterinary notes into the 7,739 SNOMED-CT codes used by the CSU Veterinary Teaching Hospital. Remarkably, fine-tuning these models on a dataset of over 246,000 veterinary visits achieved impressive results, with GatorTron leading the pack with a 75% accuracy rate. What’s even more exciting? The study showed that even smaller, less computationally demanding models achieved comparable results, opening doors for wider implementation, even in resource-constrained settings. While the fully automated coding of medical records still requires human oversight, this research paves the way for significant improvements in veterinary care and research. Imagine AI assisting veterinary staff by suggesting diagnoses in real-time, freeing up valuable time and resources. This AI-powered approach could also empower clinicians by providing them with instant coding suggestions as they write notes, streamlining documentation. Looking ahead, the researchers are focused on building trust and transparency in these AI systems. Future studies will explore methods to explain how these models arrive at their diagnoses, ensuring greater confidence in their predictions. They'll also evaluate the models' generalizability across different veterinary practices and incorporate richer data sources, like histopathology reports, to paint a more complete picture of animal health. The ultimate goal? A future where AI helps us unlock a deeper understanding of animal health, leading to improved outcomes for our furry companions and contributing to a healthier planet.