Health Professional Students' Use of Generative Artificial Intelligence During Clinical Placements: Cross-Sectional Online Survey Study

Clinical Snapshot

Key Findings

• P Value: Discipline differences significant (P=0.03), maturity trend highly significant (P<0.001)

• Effect Size: Strong association between self-perceived maturity and adoption (minimal: 9% vs high: 76%)

• Primary Outcome: 52.6% of health professional students reported using GenAI during clinical placements

• Nnt Or Sensitivity: 23.5% of users reported disclosing patient-identifying information at least once

• Confidence Interval: Midwifery students had lower uptake (OR 0.30, 95% CI 0.11-0.77)

Clinical Application

Findings suggest need for structured curricula and governance frameworks before widespread implementation Relevant to Australian health professional education but requires consideration of local regulatory frameworks (AHPRA, university policies) and may differ from French healthcare system context Health professional students in clinical training across medicine, nursing, pharmacy, and allied health

Abstract

BACKGROUND: Generative artificial intelligence (GenAI) is rapidly expanding in higher education and clinical practice. However, its use during clinical placements, where cognitive demands and responsibility for patient care increase, remains insufficiently documented. OBJECTIVE: This study aimed to characterize self-reported GenAI use during clinical placements, perceived benefits and risks, and related training and governance needs. METHODS: We conducted a cross-sectional online survey at a French university (July 17 to September 30, 2025). Eligible participants were students in medicine, pharmacy, nursing, midwifery, or physiotherapy who were currently in, or had completed within the past 18 months, a clinical placement. A 61-item questionnaire (comprising closed- and open-ended items) assessed GenAI use, task patterns, perceived benefits or risks, and training or governance needs. A composite index classified self-perceived GenAI maturity as minimal, limited, moderate, or high. Group comparisons used χ2 tests; maturity gradients used trend tests. RESULTS: A total of 388 students responded (n=308, 79.4% women), mainly nursing students (n=217, 55.9%). Overall, 204 (52.6%) students reported using GenAI during clinical placements. Use differed across disciplines (χ24=10.71; P=.03), with lower uptake in midwifery (6/23, 26%; odds ratio 0.30, 95% CI 0.11-0.77). Adoption increased markedly with self-perceived maturity (minimal: 2/22, 9% vs high: 22/29, 76%; trend P<.001). Among the 204 users, the most commonly reported uses were information retrieval (n=159, 77.9%), bibliographic search (n=152, 74.5%), and translation or rephrasing (n=145, 71.1%); patient-facing activities were less frequently reported (eg, patient-document drafting or communication preparation: n=78, 38.2%). Although most users reported never entering direct patient identifiers, 48 (23.5%) reported at least 1 disclosure of patient-identifying information, and 96 (47.1%) reported processing real medical content perceived as anonymized. The most endorsed perceived benefits among the 388 students were documentation support (n=315, 81.2%) and improved access to information (n=266, 68.5%). The most endorsed risks were dependency (n=353, 90.9%), skill erosion (n=329, 84.8%), and confidentiality breaches (n=339, 87.4%). Training needs were highest for ethics or regulatory training (294/378, 77.7%) and a best-practice clinical guide (292/373, 78.3%). CONCLUSIONS: GenAI is already used by a substantial proportion of French students in health professions during clinical placements, predominantly for information and documentation support rather than patient-facing activities. Self-perceived readiness is strongly associated with adoption. Reported disclosures and concurrent concerns about dependency, skill erosion, and confidentiality support the need for structured curricula and clear governance frameworks to enable responsible, patient-centered integration of GenAI into clinical education.

References

1. 1.Kotzki, S., Massonnet Turner, C., Gauthier, K., Minoves, M., & Vuillerme, N. (2026). Health Professional Students' Use of Generative Artificial Intelligence During Clinical Placements: Cross-Sectional Online Survey Study. JMIR Medical Education. https://doi.org/10.2196/85243