Veterinary students use AI, but training remains uneven
Bottom line
CURRENT BRIEF VERSION: Veterinary students in Spain and Portugal are already using artificial intelligence, but mostly through informal channels rather than consistent classroom training, according to a new cross-sectional study in Frontiers in Veterinary Science. The survey of 340 undergraduate and postgraduate students during the 2023–2024 academic year found that prior AI training was linked to higher self-perceived knowledge, greater use, and more positive attitudes toward AI. The study also found meaningful variation between institutions, with students at the University of Évora and CEU Cardenal Herrera University reporting stronger scores than some peers at other schools, suggesting AI readiness is developing unevenly across programs. (frontiersin.org)
Why it matters: For veterinary professionals and educators, the findings add to a growing signal that students are not waiting for formal policy or curriculum changes before adopting AI tools. That creates both an opportunity and a risk: students may arrive in clinics and workplaces comfortable with generative AI, but without shared standards for accuracy, privacy, academic integrity, or appropriate clinical use. That matters beyond the classroom, because AI is also starting to shape veterinary research and record-based workflows: a recent Veterinary Pathology scoping review found that veterinary electronic health records are often stored as unstructured free text and outlined practical ways to use natural language processing and large language models to extract usable clinical data. Recent veterinary education literature has pushed for structured AI literacy in the curriculum, while other research has cautioned that AI tools can still underperform on veterinary assessments and should be used carefully. (frontiersin.org)
What to watch: Expect more veterinary schools to move from ad hoc use toward formal AI guidance, literacy training, and course-level policies over the next academic year, especially as AI becomes more relevant not just for studying but for handling clinical information and research workflows. (eprints.gla.ac.uk)
CURRENT FULL VERSION: A new study suggests veterinary education is entering a more practical phase of the AI debate: students are already using the tools, but institutions haven't integrated them evenly. In a survey published last month in Frontiers in Veterinary Science, researchers assessed self-perceived AI knowledge, use, and attitudes among 340 veterinary students in Spain and Portugal and found that prior training and institutional context shaped how prepared students felt to use AI. (frontiersin.org)
The study comes as veterinary education, like human health education, is shifting from asking whether students use generative AI to asking how schools should teach them to use it responsibly. The authors note that AI has moved quickly into higher education, especially through tools such as ChatGPT and Gemini. A related 2026 Frontiers perspective argued that veterinary medicine still lacks a widely accepted AI curriculum, even as the profession increasingly calls for AI literacy and supervised, assistive use. (frontiersin.org)
In the Spain-Portugal survey, conducted during the 2023–2024 academic year, students who had received any prior AI training, whether self-directed, university-based, or external, reported higher self-perceived knowledge and use scores and more positive attitudes toward AI. The researchers also found statistically significant differences between institutions. Students at the University of Évora and CEU Cardenal Herrera University in Valencia reported some of the highest scores, while lower scores were reported at the Polytechnic Institute of Portalegre’s Escola Superior Agrária de Elvas, the University of Zaragoza, and the University of Santiago de Compostela in Lugo. Daily social media use showed a small but significant positive correlation with AI knowledge, use, and attitudes, which the authors interpret as a marker of broader digital engagement. (frontiersin.org)
That pattern fits with other recent veterinary education research. A separate student-perspective study in Spain found strong interest in integrating AI into veterinary training, but also notable uncertainty about clinical applications, diagnosis, treatment recommendations, and data privacy. Internationally, a larger cross-disciplinary survey of medical, dental, and veterinary faculties reported that students generally want more AI education, even though formal teaching remains limited. (pmc.ncbi.nlm.nih.gov)
Industry and student commentary points in the same direction. Portuguese trade outlet Veterinaria Atual framed the new paper as evidence that veterinary students in Portugal and Spain are already using AI in academic work while institutional integration remains unequal. In an April 2026 AAHA opinion piece, veterinary student Jeremiah Pouncy wrote that AI is already part of how some students study, synthesize notes, and generate case-based practice questions, while also arguing that student voices are often missing from the broader professional debate. Those reactions don't replace formal evidence, but they do suggest the paper is landing in a profession already negotiating real-world use, not hypothetical adoption. (veterinaria-atual.pt)
Why it matters: For veterinary professionals, the practical issue isn't simply whether AI belongs in education. It's whether new graduates will enter practice with consistent training in when to trust AI, when to verify it, and when not to use it at all. That matters for clinical reasoning, recordkeeping, communication, and data governance. A recent scoping review in Veterinary Pathology highlighted the expanding role of natural language processing and large language models in extracting and structuring veterinary electronic health record data. The review screened 5,796 papers and included 23 veterinary and 31 human studies, finding that veterinary work has more often relied on larger supervised datasets while human medicine is moving toward prompt-based LLM approaches such as GPT and LLaMA that can work with smaller annotated datasets. The authors also proposed a practical framework covering data preparation, privacy and platform choices, and prompt engineering for clinical data extraction workflows, underscoring that future veterinarians may encounter AI not just as a study aid, but as part of everyday research and record-based workflows. At the same time, research showing that AI models can underperform on veterinary examinations is a reminder that familiarity with the tools isn't the same as competence in using them safely. (frontiersin.org)
The bigger takeaway is that uneven exposure in school could translate into uneven readiness in practice. Programs that build structured AI literacy may be better positioned to teach prompt hygiene, bias awareness, privacy safeguards, and human oversight before students carry those habits into clinics. That would align with emerging school-level guidance elsewhere, including formal responsible-use documents and college policies that treat AI as permissible in some contexts, but only with transparency and boundaries. (eprints.gla.ac.uk)
What to watch: The next step is likely curricular, not technological. Watch for veterinary schools and accrediting or professional bodies to publish more explicit AI-use standards, course policies, and literacy frameworks, especially as the 2026–2027 academic year begins and student use continues to outpace formal instruction. As AI tools become more capable of turning unstructured clinical notes into analyzable data, schools may also face pressure to teach not just chatbot etiquette, but the basics of privacy-conscious clinical data handling and AI-assisted documentation workflows. (frontiersin.org)