Application and ethical implication of generative artificial intelligence in medical education: a cross-sectional study among critical care academic physicians in China

BMC Med Educ. 2025 Aug 29;25(1):1225. doi: 10.1186/s12909-025-07825-0.

Abstract

Background: This study explored the use of Generative Artificial Intelligence (GAI) by Chinese academic physicians in clinical teaching activities within standardized residency training (TA-SRT) in critical care settings, and their awareness of ethical issues associated with GAI. The findings of this study will guide the rationale for GAI applications and sustainable integration into medical education.

Methods: This nationwide cross-sectional study utilized a self-administered questionnaire distributed via the Wenjuanxing platform to attending and higher-ranking physicians in critical care medicine departments across China who were involved in TA-SRT. Data were collected from December 2024 to January 2025.

Results: Among 456 enrolled academic physicians, 64.7% used GAI in clinical medical work and 33.1% used it in TA-SRT (GAI-User). GAI-Users reported higher training levels of GAI (p < 0.0001) and were more supportive of increased GAI training (p = 0.0368 and 0.0064, respectively). They also showed greater trust in the GAI judgment (p < 0.0001) and optimism regarding its application. Top GAI uses included querying teaching content (79.5%) and creating teaching materials (65.6%). Over 50% of physicians were aware of ethical issues, with “over-reliance” (85.7%) and “data privacy” (84.0%) being major concerns. Most agreed ethics should be part of GAI education (94.3%).

Conclusions: Although awareness of GAI is growing among Chinese academic physicians in critical care, its practical use in clinical teaching remains limited. These academic physicians acknowledge GAI’s potential to enhance clinical teaching quality and promote educational equity. To foster broader adoption, cultivating the clinical application skills of GAI is essential to lay the foundation for its educational use. Additionally, GAI training programs should integrate both practical application skills and ethical education, which are widely supported by academic physicians, to guide the rationale and sustainable integration of GAI into medical education.

Supplementary Information: The online version contains supplementary material available at 10.1186/s12909-025-07825-0.

Keywords: Critical care medicine; Ethics; Generative artificial intelligence; Medical education; Standardized residency training.