Leveraging artificial intelligence for ICU pharmacotherapy training: Supporting learning while preserving clinical judgment
Artificial intelligence (AI) is increasingly embedded in critical care practice at a time when pharmacotherapeutic decisions are becoming more complex, data-intensive, and time-sensitive. For medical residents and fellows, this creates both educational opportunities and challenges, as AI may support learning but must not replace the development of independent clinical judgment. This perspective examines potential applications of AI in ICU pharmacotherapy training, including support for just-in-time learning, case-based practice, and bedside decision support. Key limitations and risks, such as misinformation, cognitive offloading, deskilling, privacy and data security, and bias in model outputs, are discussed along with possible mitigation strategies. Future training should prioritize building competencies related to AI-assisted clinical decision-making, including critical appraisal of AI-generated outputs and communication of AI-informed decisions within interprofessional teams. AI-supported pharmacotherapy training should be guided by clear boundaries around use and continued emphasis on building foundational pharmacotherapy knowledge, patient-specific reasoning, and clinical judgment.
Authors
- Eunice S. Kim
- Jessica D. Workum (ORCID: https://orcid.org/0000-0002-4134-0242)
- Ji-Yeon Kim
- W. Anthony Hawkins
Institutions
- University of Georgia (US)
- Augusta University (US)
- Winchester Medical Center (US)
- Elisabeth-TweeSteden Ziekenhuis (NL)
Publication Details
- Journal
- Journal of Critical Care
- Published
- 2026-11-13
- DOI
- https://doi.org/10.1016/j.jcrc.2026.155820
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00