Most health science students are already using artificial intelligence tools in their studies, but a new survey suggests their institutions have largely failed to keep pace. The findings paint a picture of rapid, largely self-directed adoption of AI alongside persistent gaps in formal instruction and widespread ethical unease.
The scale of everyday use stands out. Ninety percent of participants, 279 students, reported using ChatGPT, one of the most widely known generative AI chatbots. Yet formal education on these tools has not followed: 82.6 percent of respondents said they had received no instruction on AI at all. Students described only moderate familiarity with the technology, with a mean knowledge score of 3.08 on the study’s scale, with a standard deviation of 1.06. That combination, near-universal use paired with minimal structured teaching, suggests students are learning to work with AI largely on their own.
The study also examined how AI use related to academic performance. Students who used AI tools more frequently reported higher grade point averages, and the association was statistically significant in an analysis of variance, with F = 15.56. The authors emphasize, however, that the cross-sectional design of the study cannot establish that AI use causes better academic outcomes. It is equally possible that stronger students gravitate toward these tools, or that other factors influence both usage and grades. Self-reported measures add a further layer of caution, since students’ own accounts of their study efficiency and comprehension may not align with objective performance.
Despite those caveats, the self-reported academic picture was broadly positive. Participants linked AI tools to greater study efficiency and better comprehension of their coursework. For students juggling dense curricula across medicine, nursing, and allied health disciplines, the appeal of tools that can summarize complex material or help draft study notes is easy to understand. The survey results indicate that this appeal translates into heavy real-world use, even in the absence of institutional guidance on how to use such tools well.
Ethical concerns emerged as a major theme. Sixty-one percent of participants reported ethical worries about AI in their education. Those concerns were not evenly distributed: they varied significantly by gender, with p = 0.0015, and by academic program, with p = 0.0114. The variation across programs suggests that students in different health disciplines encounter AI differently, whether through differing coursework, clinical exposure, or expectations about professional practice, and that their ethical reservations reflect those distinct experiences.
The authors argue that the combination of high use, low formal training, and prevalent ethical concern points to a need for structured curriculum change. Their recommendation for practice is that faculty implement what they call proactive AI training, designed in advance of problems rather than in reaction to them. Such training, they suggest, should address automation bias, the tendency to over-trust outputs from automated systems, and build ethical literacy so that students can critically evaluate AI-generated content rather than accept it uncritically.
The stakes extend beyond individual study habits. Health professions graduates will enter fields where AI tools are increasingly present in clinical and research settings, and equitable competence across disciplines matters for patient care. The finding that ethical concern differs significantly by program hints that some student groups may be better prepared than others to navigate these tools, a disparity the authors link to the need to ensure equitable technological competency across healthcare disciplines.
It documents associations, such as the link between usage frequency and GPA, and patterns, such as the gender and program differences in ethical concern, but it cannot determine direction of causation or long-term effects. The authors state this limitation explicitly in their conclusion, noting that the cross-sectional design does not establish causality between AI use and academic achievement.
What the survey does establish is a clear baseline. AI tools are already woven into the daily study routines of the vast majority of these health science students, formal instruction has not kept up, and a majority carry ethical reservations that vary across the student body. The authors’ conclusion is that these disparities in formal training and significant ethical issues underscore the need for a structured curriculum, one that treats AI literacy as a teachable competency rather than leaving students to improvise. As generative AI continues to spread through education and healthcare alike, studies like this one offer institutions a starting point for deciding what, and how, to teach about it.
Subject of Research: AI use, self-reported academic effects, and ethical concerns among health science students
Article Title: Bridging technology and pedagogy: a cross-sectional analysis of AI use, self-reported academic effects, and ethical concerns among students across diverse health science curricula
Article References: Alshahrani, A. M., Ibrahim, I. A. E., Jibo, A., Alshahrani, A. F. S., Alshahrani, A. F. A., Alshahrani, K. H. S., Alqarni, G. A. M., Alghamdi, M. A. S., Almarhabi, A. M., Almuawi, M. T. M., Alharthi, S. H., & Albishi, D. M. K. (2026). Bridging technology and pedagogy: a cross-sectional analysis of AI use, self-reported academic effects, and ethical concerns among students across diverse health science curricula. BMC Medical Education. https://doi.org/10.1186/s12909-026-10645-5
Image Credits: AI Generated
DOI: 10.1186/s12909-026-10645-5
Keywords: artificial intelligence, health professions education, ChatGPT, GPA, ethical concerns, cross-sectional study, medical education, survey, curriculum, training gap, Bridging, technology
