Artificial intelligence is upon us and likely will forever change the way we interact with learning and education. Despite this reality, educational institutions seem to fall into either of two camps. One camp seems loath to acknowledge that AI exists. A faculty member who helps with curriculum development at one medical school recently shared, “We don’t know what to do about AI. Do we act like it’s not there, or do we acknowledge it?”
The other camp embraces AI and encourages students to employ AI resources, such as ChatGPT. Given the possibility of plagiarism or simply allowing students to bypass any learning whatsoever, it’s understandable why medical schools and other institutions might be conflicted about AI. But even though AI is no replacement for a novel idea or human thought or in medicine placing one’s hands on patients, AI has value in medical education–and not just a little bit. AI may be used as a supplement, resource, or aid when we are learning, teaching, or creating something new.
Just how could medical schools and medical students use AI to assist in educating students? Prior articles have suggested how medical students can utilize chatbots, like ChatGPT, as online tutors to help answer questions or to create quizzes to test their knowledge. For example, bots like ChatGPT can help compare differences in diagnoses, treatments, or procedures that students may be confused about. Those same AI sites can offer a personalized learning experience that schools ought to acknowledge or promote. NYU Grossman School of Medicine has run with this idea and has fully embraced the idea of precision learning from AI by incorporating a “precision education” tool. Each NYU medical student is offered a personalized medical education, with an AI algorithm tailoring subject matter and content format.
In the research space, AI can also be invaluable in medical education. For example, faculty and students alike can also utilize AI to help create data analysis plans, code for various computer languages and scan literature. An online website called Elicit lets users pose a question and then, through AI, scans the internet to find papers and synthesizes their findings into a summary.
Outside of the student experience, professors may also use AI to create lecture outlines and predict the questions that students are most likely to have about certain material. Additionally, professors and faculty must be able to set standards and address the use of AI in the classroom. If they don’t, students may misunderstand the expectations for AI and when or if its use is permissible.
In our own experience, we have used AI to create study guides for courses, create outlines for lectures and book chapters, analyze CVs, and write initial drafts of promotion letters for fellow faculty members. We are certain that the uses of AI that will further simplify our work and assist in medical learning will become clearer and only be seen as greater assets going forward.
Medical schools already offer courses on a wide range of learning and research topics, such as “best study” practices or how to conduct a literature review. Going forward, AI-based tools should be included in these lectures and within the list of online resources for student learning and research. Additionally, schools should teach students what to watch out for when using AI, like bias or flat-out false information and/or non-existent references. Teaching how to use AI in one’s learning promotes a more prepared generation for future technological innovation. This approach may complement courses that explore innovation and AI in medicine.
To those who are hesitant to incorporate AI into education even after reading about NYUs approach and our own ideas, we encourage them to look at how AI has improved other aspects of medicine. From image analysis in radiology and pathology to quick retrieval of medical information and tracking infectious disease outbreaks, this technology has created greater efficiency in health care. Other studies have found that AI can help reduce racial disparities in health care, with one investigation finding that AI better predicted pain from X-rays for underserved patients when compared to radiologists. The technology can be used for good, including in education.
To illustrate that we are not just talking about the potential values of AI in medical school education, in thinking about writing this essay, we asked ChatGPT, “how can artificial intelligence be used to teach medicine and enhance learning in medical schools.” The answers ChatGPT provided included personalized learning, virtual patients, data analysis and research, smart tutors, and educating students about the limitations, biases, and potential risks of AI tools.
As anyone ought to do when using AI, we analyzed ChatGPT’s response, and ultimately–although this might not always be the case–we agree with its recommendations. Therefore, given that we are intent on practicing what we’re now preaching, we couldn’t have written our piece without emphasizing those elements, among others.
Amelia Mercado is a medical student. J. Wesley Boyd is a psychiatrist.
