How artificial intelligence can be used in veterinary medicine and education

Photo: snowing12/Adobe Stock

Candice Chu, DVM, PhD, an assistant professor at the Texas A&M College of Veterinary Medicine and Biomedical Sciences (VMBS) in College Station, is discovering new uses for artificial intelligence (AI) tools in veterinary clinics and classrooms.1 With her research, Chu has developed a new AI-based study tool for veterinary students. The tool, VetClinPathGPT, is based on ChatGPT technology.¹

“AI is a powerful assistant that can help clinicians, educators and students reduce the time they spend on repetitive work,” Chu said in a news release.¹ “By reducing repetitive tasks, individuals can spend their time on the most important parts of their work, like completing assignments or helping students and patients.”

AI in the Classroom

Chu has expressed that it is both her desire, and her expectation, that AI will become a regular facet of the veterinary classroom in the future.¹ VetClinPathGPT is designed to assist students as they learn clinical pathobiology, focusing on disease diagnosis, treatment, and prevention. Students can ask the AI questions about terms and concepts that they are unsure of, and educators can use AI tools to reduce their time spent on repetitive tasks, like writing exam questions.¹

“Imagine having a microscope connected to a camera with AI capability so students can look down at their slides and essentially have a clinical pathologist there to explain what they’re seeing,” Chu said of AI’s upside as a learning tool.¹ “They could ask the AI, ‘What’s that cell?’ and it could tell them not only the type, but also how to identify it. Veterinarians are in high demand, and it’s a good thing to have lots of students in the classroom. But 1 instructor cannot work with every student individually at the same time, so in my ideal world, AI would be able to help with that.”

Chu also explained the role that VetClinPathGPT can play as a study partner. Students can upload their course readings, and the AI can ask them relevant study questions based on the text. “Once you read one passage, you can test your knowledge or use the AI to help you prepare for the exam,” Chu said.¹

AI in the Clinic

Chu predicts that AI tools will continue to serve veterinary professionals well past their time in a classroom. The tools can assist with tasks like record keeping that traditionally take away from time otherwise spent working with patients.¹

“It would be very helpful to have a tool that could go through medical record and pull examples from cases,” Chu said of AI’s potential use in clinical settings.¹ “Imagine having a tool that could fill out medical records while you’re talking with a patient’s owner.”

Any cause for concern

A primary drawback regarding the use of AI in such avenues has to do with privacy concerns and ensuring that these tools don’t compromise the privacy of confidential medical records. Chu expressed shared concerns over the fact that, for AI tools to read patient medical records, those records would need to be shared with developers, which would violate patient privacy laws. She is confident, though, that it’s only a matter of time before someone develops AI tools that satisfy patient privacy protocols, referencing pre-existing private versions of ChatGPT, or the possibility of an AI tool that exists only locally on a clinic’s computer or network.¹

Another major concern addresses the accuracy of AI-provided information. “Artificial hallucination” refers to the generation of implausible but confident responses by an AI tool.² Chu cited a 2023 article that covers these hallucinations, noting that ChatGPT sometimes fabricates references with incoherent PubMed IDs.³

As large language models (LLMs), tools like ChatGPT are trained on accessible online data and are refined through conversations with users. This, as Chu points out, can also raise some concerns regarding copyright infringement and privacy violations. She also notes that some may be concerned by the fact that the FDA has not yet authorized any medical devices that use generative AI or LLMs.²

“The key to getting high-quality, relevant answers to your prompts is to make sure that the information uploaded to the AI is reliable,” Chu said, acknowledging concerns that some may have regarding the accuracy of AI-provided information.¹ “For example, VetClinPathGPT doesn’t use just any information from the internet, it uses the ‘eClinPath’ website, an online textbook developed by the Cornell University College of Veterinary Medicine [in Ithaca, New York].”

An advocate for its implementation, Chu sees AI as the ideal complement to humans in veterinary medicine, not as a replacement. “Neither are perfect on their own, so bringing them together increases our ability to diagnose, teach, and learn” Chu said.¹ “One thing I tell people is not to worry that AI will replace you; it’s the person who knows how to use AI who will replace you. I think that in the future, knowing how to incorporate this technology will be a basic requirement to be an efficient and competitive veterinary educator in the job market.”

In addition to her research and role as an assistant professor at VMBS, Chu announced a partnership with VMBS’ Gastrointestinal Laboratory and the Texas A&M Institute of Data Science on a machine learning project to see if AI can help diagnose acute pancreatis in dogs. This fall, Chu will deliver a keynote lecture this fall at the American College of Veterinary Radiology Scientific Conference, and is also scheduled to attend the Texas Taiwanese Biotechnology Association Symposium and the American College of Veterinary Pathologists Annual Meeting, where she is expected to share her research.¹

References

1. Texas A&M Researcher Explores Benefits of AI In Veterinary Medicine, Education. News release. Texas A&M College of Veterinary Medicine and Biomedical Sciences Communications. August 20, 2024. Accessed August 21, 2024. https://vetmed.tamu.edu/news/press-releases/ai-in-veterinary-medicine/
2. Chu C. ChatGPT in veterinary medicine: a practical guidance of generative artificial intelligence in clinics, education, and research. Front. Vet. Sci. 2024;11. https://doi.org/10.3389/fvets.2024.1395934
3. Alkaissi H, McFarlane SI. Artificial Hallucinations in ChatGPT: Implications in Scientific Writing. Cureus. 2023;15(2):e35179. Published 2023 Feb 19. doi:10.7759/cureus.35179