[This blog entry was created using AI tools as part of an exploratory process. The initial step involved identifying relevant articles using Open Evidence (OpenEvidence). After filtering the results, the writer selected the four most salient articles for deeper analysis. These were then input into ChatGPT with instructions to generate a 500-word blog post tailored to an audience of medical educators across disciplines. Several rounds of revisions followed, involving both AI-assisted editing and human review, including careful cross-checking of references. This process illustrates one example of how leveraging the strengths of various AI tools can enhance academic and professional communication.]
Artificial intelligence (AI) isn’t coming to healthcare—it’s already here. From diagnostic support to personalized treatment planning, AI is reshaping how clinicians work. Yet, as the tools grow smarter, are we making sure our future doctors are ready to use them wisely?
For many medical schools, integrating AI into the curriculum still feels like a futuristic challenge, reserved for tech-forward institutions or students with a coding background. But that mindset may be doing our learners—and their future patients—a disservice.
As Lee et al. (2021) noted in their scoping review, most current efforts to teach AI in undergraduate medical education are fragmented and inconsistent. Students get glimpses of AI—perhaps in a radiology elective or a data science lecture—but rarely develop a foundational understanding of how it fits into clinical reasoning, ethical decision-making, or patient care.
Fortunately, the conversation is shifting. Gordon et al. (2024), in their comprehensive BEME review, advocate for a structured, competency-based approach to AI education. Rather than teaching “about technology,” they propose helping students become critical consumers of AI—able to evaluate algorithms, question biases, and make informed clinical choices in AI-augmented environments.
What should this look like in practice? Çalışkan et al. (2022) provide a helpful starting point. Through a Delphi study, they identified core competencies that medical students should acquire, including data literacy, understanding the limitations of AI, and ethical awareness. These aren’t just technical skills—they’re critical thinking tools.
This shift isn’t about turning every med student into a programmer. It’s about preparing them to ask smart questions: Can I trust this algorithm’s output? Is it transparent? Could it reinforce existing healthcare inequities?
Engaging students in these conversations is not just possible—it’s essential. As Chen et al. (2019) emphasized, students respond well to case-based and clinical examples where AI tools are already in use. Imagine a cardiology module that includes AI-based EKG interpretation, or an ethics seminar that tackles algorithmic bias in patient triage.
Wartman and Combs (2018) take it a step further, arguing that AI should catalyze a reimagining of medical education itself. In a world where machines can recall every clinical guideline, the value of a physician lies in synthesis, judgment, and humanity. Our curricula and means of evaluation should evolve accordingly.
And AI isn’t just the content—it can also be the method. Tools that offer AI-powered feedback (Aquifer, 2025) or assist in curriculum mapping (Raskob et al., 2025) are already helping educators personalize learning and identify gaps.
Medical educators don’t need to become data scientists—but we do need to lead this transformation. Let’s give our students the tools to engage with AI thoughtfully, challenge it where needed, and use it to deliver more equitable, effective care. Because the future of medicine isn’t just high-tech—it’s human-led, AI-enabled.
What do you think?
- Where in your current curriculum could AI naturally be introduced or integrated?
Think about existing modules (e.g., radiology, pathology, ethics) where AI tools are already being used in clinical practice. - What competencies do you think are most important for students to develop regarding AI in medicine?
How might you prioritize ethical reasoning, data literacy, or critical evaluation of AI tools? - How prepared do you feel—individually or institutionally—to teach about AI in medical education?
What support, training, or partnerships would help you feel more confident? - How can AI be used not just as content, but as a tool for teaching and assessment in your program?
Have you considered ways to use AI for formative feedback, curriculum mapping, or adaptive learning? - What potential risks or unintended consequences should we be mindful of as we integrate AI into medical education?
How do we ensure equity, transparency, and student-centered learning while adopting new technologies?
References
Aquifer (2025). AI-Powered Feedback: Enhancing Clinical Reasoning in Medical Education. Available at https://aquifer.org/blog/ai-powered-feedback-enhancing-clinical-reasoning-in-medical-education/. Accessed June 17, 2025.
Çalışkan, S. A., Demir, K., & Karaca, O. (2022). Artificial intelligence in medical education curriculum: An e-Delphi study for competencies. PLOS ONE, 17(7), e0271872. https://doi.org/10.1371/journal.pone.0271872
Chen, J. H., Asch, S. M., & Wren, S. M. (2019). Artificial intelligence in medical education: A review. JMIR Medical Education, 5(1), e13930. https://doi.org/10.2196/13930
Gordon, M., Daniel, M., Ajiboye, A., Atkinson, H., Downer, J., & Parise, A. (2024). A scoping review of artificial intelligence in medical education: BEME Guide No. 84. Medical Teacher, 46(4), 446–470. https://doi.org/10.1080/0142159X.2024.2314198
Lee, J., Wu, A. S., Li, D., & Kulasegaram, K. M. (2021). Artificial intelligence in undergraduate medical education: A scoping review. Academic Medicine, 96(11S), S62–S70. https://doi.org/10.1097/ACM.0000000000004291
Raskob, K., Duman, H., Kinder, J., Lee, J., Wilson, J., & Segerson, K. (2025). Twelve tips to harness the power of AI for curriculum mapping. Medical Teacher, 1–10. https://doi.org/10.1080/0142159X.2025.2513427
Wartman, S. A., & Combs, C. D. (2018). Reimagining medical education in the age of AI. Academic Medicine, 93(8), 1107–1109. https://doi.org/10.1097/ACM.0000000000002111
Author: Silka Patel, M.D., MPH; Association of Professors of Gynecology and Obstetrics
Bill Fuller says:Reply
June 16, 2025 at 2:09 pm
I commend the authors for highlighting the central issues in training students to be intelligent and critical users of AI. I take two issues with the piece:
-It draws no distinction between machine learning (e.g. AI assisted ECG interpretation) and Generative AI. GenAI being used to replace primary source or point of care reference review is a unique problem in the way that it changes the cognitive work of medicine, and requires an approach distinct from other AI tools. There are commonalities in the necessity of understanding the algorithms, but the shortcut that GenAI in particular offers to avoid critical review of primary literature has the potential to be deeply detrimental to learning in a way that is not applicable to other machine learning tools.
-Being written primarily by AI, the piece has all of its bad habits. The GPT em dash obsession—a notable hallmark of AI writing that many professors will recognize—and the failure to appreciate context result in a piece that does not appear to know its audience, informal to the point of condescending. The introductory sentence, for example is eyerollingly obvious, and I doubt the authors would have started a piece for medical educators this way of their own initiative.
I do not say this to criticize the authors themselves! I find it tremendously valuable to see what human-AI collaboration generates, and I hope my second comment can add to that conversation and highlight specific limitations of GenAI tools. Learners should be taught how the act of writing is different than the act of editing both in its cognitive effects and in the product produced, just as the act of prompting and reading an AI literature summary is cognitively different than the act of personally performing critical analysis. Becoming expert at both is necessary.
JM says:Reply
June 16, 2025 at 2:10 pm
Having trouble finding the Ellaway and Wang articles. The Ellaway DOI does not come up and the Wang DOI is a different article. Manual searches are not leading to results either. Are these references hallucinated?
acegbeck says:Reply
June 17, 2025 at 10:00 am
Thank you for noticing that. The references were updated as those were hallucinations…which in some ways speaks to the hazards of AI generating text.
Deb Simpson says:Reply
June 16, 2025 at 2:59 pm
Nice example – and shows opportunities to use AI. As noted – authors were still the humans in the room – so wonder how they chose the articles they did? What was “time frame” as AI lit in med ed is exponentially increasing. Surprised not to have anything on cognitive deskilling as a concern/citation…. as its not whether to use or not – it’s about what we expected to learn (and faculty). To keep up – I lean on the Josiah Macy Foundation’s recent work in this area.. https://journals.lww.com/academicmedicine/abstract/9900/macy_foundation_innovation_report_part_i__current.1185.aspx
And Academic Medicine’s collection on AI. https://journals.lww.com/academicmedicine/pages/collectiondetails.aspx?TopicalCollectionId=82
Q4 Re: competencies – Right know I prefer the broader view of digital competencies- Car J, Ong QC, Fox TE, Leightley D, et al. The digital health competencies in medical education framework: an international consensus statement based on a Delphi study. JAMA network open. 2025 Jan 2;8(1):e2453131-.