Augmented intelligence (AI)—commonly called artificial intelligence—is quickly becoming part of how medical students study, review lectures and look up unfamiliar concepts. A 2025 study of medical students reported that a significant majority of respondents used ChatGPT to enhance their learning experience.
That takeaway, using only a single platform as a reference point, illustrates that AI tools will have a role in how future physicians learn. But how do AI tools need to be harnessed, and which use cases are most reliable?
Two resident physicians who have studied the use of large language models in undergraduate medical education offered insight on the do’s and don’ts of AI use for medical students.
Don’t: rely on unharnessed AI
New research suggests the biggest benefit may not be higher test scores. Instead, AI may be most useful when it helps students manage the flood of information that comes with medical school—especially in the preclinical years—by turning dense content into clearer, more usable study tools.
For a recent study published in the academic journal BMC Medical Education researchers at the Warren Alpert Medical School of Brown University created AI-generated lecture summaries and Anki flashcards for first-year medical students in genetics and pharmacology using a structured prompt-and-review process.
Researchers did not examine freeform chatbot use, however. Instead, they evaluated AI-generated summaries and flashcards that were produced through a structured process and put into a validated tool for which prompts were tested, outputs were reviewed for hallucinations and learning-objective coverage. Then the best-performing approach was used to create study materials for medical students.
The study authors’ contention is that these types of guardrails are necessary to ensure accuracy. Simply giving medical students raw ChatGPT or Claude access often doesn't help. Poorly calibrated content introduces misinformation.
“When we see undifferentiated use of AI—AI that's not controlled—then there's not always clear learning outcomes,” said AMA member Jay Khurana, MD, the paper’s lead author and a neurology resident in Connecticut. “If students are using AI in a structured way—the sort of application where it’s controlled, medically validated—that is where our study finds some improvements in learning outcomes and substantial improvements in time savings.”
From AI implementation to digital health adoption and EHR usability, the AMA is fighting to make technology work for physicians, ensuring that it is an asset to doctors. That includes the AMA Center for Digital Health and AI, which works to ensure physicians help shape how AI is developed, implemented and regulated across the healthcare system, with patient safety and physician-led care remaining at the center of those efforts.
Don’t: abandon critical thought
AI can often give students a quick, correct answer. But using AI often means students are bypassing the critical thinking that is required to grow as a learner.
- In this AMA news article, medical students will get valuable information regarding:
- How unharnessed AI can introduce misinformation or miss key learning objectives.
- The risk of deskilling when relying on AI
- How medical students can use AI for literature review, research and finding information.
- Why AI works better as a tool for active recall than for passive studying.
- The time savings of using validate AI tools to create study materials.
- How prompt wording and platform choice can shape the quality and accuracy of AI-generated study materials.
