Comparison of Generative Artificial Intelligence and Student-Generated Veterinary Handouts

Generative artificial intelligence (gAI) is becoming increasingly prevalent in our daily lives. Students across multiple disciplines are using gAI for writing assessments and completing projects. This leads to the question: could gAI platforms perform similarly, worse, or better than veterinary students when asked to create discharge handouts for select veterinary neurological conditions? A total of 24 professionals and educators graded handouts based on content, clarity, client education, empathy, and professionalism for two canine neurological conditions (seizures and intervertebral disc disease). Each condition had handouts created by a high-performing student, a random/unknown student whose work was deemed to represent the average student at our institution, ChatGPT 4.0, and Google Bard. The high-performing student's handout scored higher in several categories compared with AI-generated handouts and scored statistically higher overall. Specifically, the high-performing student's seizure handout scored significantly higher in accuracy/completeness and client education than the Bard handout. Both student handouts scored significantly higher for empathy and client support than the AI tools. For the intervertebral disc disease handouts, the AI-generated handouts scored higher in clarity and organization than the random student handout, with the high-performing student's handout scoring higher in empathy and client support over the Bard-generated handout. Upon the conclusion of grading, reviewers completed a survey asking them to guess the authorship of each handout. Veterinary educators and professionals could not distinguish between gAI- and student-developed client handouts. However, the findings suggest that students have the potential to outperform current gAI technology in multiple areas, including conveying empathy.