The integration of artificial intelligence (AI) in higher education is an emerging and rapidly evolving field with the potential to transform the learning process. AI tools can facilitate access to vast amounts of information and provide new perspectives on problem-solving. However, their effectiveness in educational settings, particularly in complex disciplines like veterinary microbiology, remains an area of active investigation. This study explores students' perceptions about the use of generative AI and examines whether its use can help veterinary students to accurately answer questions about clinical cases and solve them. 

The study involved 68 students enrolled in the “Biological Agents of Veterinary Interest” course, which is part of the Veterinary Medicine degree program at the University CEU Cardenal Herrera (Valencia, Spain). To assess the effectiveness of AI use, six clinical cases were designed and presented to students across three workshop sessions. In each session, students were divided into two subgroups: one subgroup asked the IA to solve the first three cases and in the remaining three cases, students solved the cases without using AI but using traditional study methods based on academic resources and critical thinking. Meanwhile the other subgroup followed the opposite approach. This crossover design ensured that every student asked AI to solve three cases and solved without AI the other three. The students were finally invited to complete a voluntary questionnaire after solving the cases. This questionnaire aimed to gather students' perception about the use of AI and to compare the accuracy of students' responses whether the clinical cases were resolved by AI or not.

A total of 56 students completed the questionnaire, yielding a high response rate of 82%. The results revealed that the majority of students recognized various advantages of AI, reporting that it enabled them to find information more quickly, suggested novel diagnostic approaches, and increased their confidence in their clinical decision-making. However, student responses demonstrated higher accuracy than AI responses. Specifically, 71.4% of the responses were correct when students relied on conventional learning, compared to 60.6% when AI tools were used. 

Interestingly, students’ preference aligned with these findings, as 83% of participants indicated that studying clinical cases without AI was the most effective method for knowledge acquisition, suggesting students recognize the importance of retrieval for effective memorization. While AI provided valuable support in terms of efficiency and access to diverse sources of information, students stated that studying without AI assistance would be more effective for deep learning and mastering concepts. 

These findings suggest that while AI can enhance certain aspects of case solving, such as speeding up search for information, providing innovative diagnostic ideas, and increasing students' confidence in their diagnostic decisions, traditional methods like review of reliable academic resources and critical thinking may still be more effective compared to AI-assisted approaches. AI should therefore be integrated as a supplementary tool rather than a primary method, supporting but not replacing conventional study techniques. Future research should explore how to optimize the integration of AI in veterinary curricula, ensuring that it complements rather than detracts from core learning objectives.

Keywords: Artificial intelligence, Veterinary education, Microbiology, Clinical cases.