Artificial Intelligence and Clinical Reasoning: An update Based on PubMed
Introduction
Artificial Intelligence (AI) is a evolving tool in the application to Clinical Medicine, but its incorporation into the everyday life of clinical practitioners is under evaluation. This relatively new field has shown its power in facilitating administrative and bureaucratic tasks, but the real contribution to clinical medicine is still under scrutiny. Ethical, biases and confidentiality, must be regulated in practical terms and globally.
Methods
We searched in PubMed under the terms “Artifici6al Intelligence” AND “Clinical Reasoning”. We obtained 26 papers about the subject in the last year. We selected four of these papers because of the more direct association with our search, and we will expose here in a summary form the results06. Two of these four papers have free access for reading.
Results
Here, we show the identification of the four papers, and we will comment on some important aspects later.
- Artificial Intelligence and deskilling in medicine. Monteith S, et al. Br J Psychiatry 2026 Jan(8): 1-3. This article deals with the sense of overreliance of physicians in the AI, accepting the information without criticism, showing a possible landscape of a loss of clinical reasoning and clinical judgment.
. AI-induced never-skilling in medical education. Yuhe Ke, et al. Nat Med 2026; Jun; 32(6): 1997-2006.
This paper insists on the same possible danger for the acquisition of a proper clinical reasoning in trainees who rely on AI during clinical learning, never reaching a proper clinical method. The authors differentiate two other concepts, deskilling, the situation of a loss of clinical expertise, and mis-skilling, also associated with trainees when they accept without confirmation the information from the AI, with the possibility of incorporating wrong clinical knowledge.
. Artificial intelligence in undergraduate medical education: an updated scoping review. Simoni J, et al. BMC Medical Education. https://doi.org/10.1186/s12909-025-08188-2.
They made a literature research in PubMed in 2024, and other databases.
310 publications were included in the review. They confirm an increase in the use of AI in medical education, but they don´t find a clear path to evaluate the impact on critical thinking and clinical reasoning.
. Comparing Artificial Intelligence to physicians´ competences in the domain of clinical reasoning:Mlika M, et al. Journal of Medical Education and Curricular Development. DOI: 10.1177/23821205261463619
The authors performed a meta- analysis, and they used 1609 clinical vignettes to compare different applications of Large Language Model to human clinical reasoning. No significant differences were detected, but there was a high component of heterogeneity inside the human group.
Conclusion
These four paper are an example of the general landscape for AI in relation to medical practice and clinical reasoning. In general the summary information that we can recover until now is the following:
AI is very interesting in Medical Education, but the difficulty is to organize the inclusion inside the clinical curriculum.
AI is a strong help in clinical reasoning, but most of the comparisons don´t show significant differences between clinicians and AI. The reason for that is the lack of a proper method of comparison, and the heterogeneity of the group of comparison. What is clear is the AI can help at least in terms of human memory, but maybe the human is better trained to detect common diseases with an atypical presentation.
AI can have also unpredictable consequences, such as to influence a desskilling clinical process, trying to choose a faster and easier path in the mind of the practical clinician.
Author:
Dr. Lorenzo Alonso Carrión
FORO OSLER
Donate: Paypal @Forooslerassociation
