Everyone is using ChatGPT to breeze through college. What does this mean for medical school admissions?
With AI further proving that our pre-med system is outdated, how can we continue to select the best to enter the medical field?
College, as we have known it for hundreds of years, is over.
This was the conclusion I came to after reading James Walsh’s fantastic recent article in New York Magazine, which details the rampant rise of AI use by students in college across nearly every discipline. The story is filled with narratives and quotes of several current college students, several of which I have included here.
Just two months after ChatGPT was released in November 2022, a survey of 1,000 college students found that nearly 90 percent had used the chatbot to help with homework assignments. And that was two years ago. Now, ChatGPT is the 5th most visited website in the world (higher than WhatsApp and X).
Readers of my Substack will also know that AI has been rapidly incorporated into many industries, including healthcare. You can check out some of my other thoughts on that here.
AI has become synonymous with education of any sort, being utilized to do everything from taking notes, creating study guides and practice tests, summarizing novels and textbooks, and writing entire essays from scratch. In the STEM field in particular, students are completely automating their research and data analyses and flying through dense coding assignments. Despite being just a few years out of college myself, I can only imagine how different the landscape must be for current students with this tool at their disposal. I myself have used ChatGPT in my research work helping me code statistical analysis in RStudio and to scan my methodology for potential limitations. I can personally attest that its capabilities are extremely impressive and that it could have the potential to revolutionize education for the better.
However, there is no denying that some of what it has caused is a disregard for academic honesty. Newer surveys have found that almost half of college students admitted using ChatGPT for an at-home test or quiz and almost a quarter using it to craft entire essays. I personally think the real number is much higher. One aspect of the NYMag article I particularly found thoughtful was the fact that in the years prior to the release of ChatGPT, most college students had already undergone years of rapid transition to online schooling, largely due to the effects of the COVID pandemic. Cheating in school had already reached somewhat of a zenith as most assignments, classes, and examinations had become entirely remote and unsupervised. In this sense, ChatGPT coming to the table made an already existing problem even worse. It is clear to me now that cheating on schoolwork is no longer an activity done by a few struggling students in a backroom trying to keep a low profile - it is done brazenly and by a majority of students. I truly believe that very little learning is happening in classrooms today and that this will not change until our schools adapt.
Many have attempted to fight this by outright banning the use of AI and implementing AI-detection software. However, these have wildly different success rates and studies have shown that they flag more false positives than they identify true positives. James Walsh himself put the Book of Genesis into one of these AI-detectors and it came back as 93% AI-generated.
I strongly disagree with this approach - there is simply no way to ban our way out of the AI era. Others schools have transitioned to individual 10-15 minute oral or discussion-based exams actually encouraging the use of AI to prepare - I think this is a creative idea that is a step in the right direction.
But what does this mean for pursing medicine, which remains one of the most competitive paths a student can take after college? It is no secret that admission has become more and more selective each year, with one of the most important parts of the application being college grades, particularly in pre-med coursework. With the increasing abundance of ChatGPT in classrooms and widespread cheating that will likely go undetected by professors, how can we adapt to ensure that we are still selecting the best to follow us into medical school?
This trend is only the latest that has shown that our pre-med system is outdated, not data-driven, and in need of a serious revamp. I have written about this extensively, arguing that the required college coursework is largely irrelevant and unjustifiably constrains students. Just as colleges have to reimagine a new system of educating students, medical schools should also take the chance to revisit what they look for in competitive applicants.
In my view, this involves a new philosophy - evaluate applicants as people, not just as students.
Pre-med coursework requirements should be largely reduced to the essentials (e.g. just biology and biochemistry) and students should be given free rein to study what they want in college. While good grades and test scores are expected of any applicant, weight should be placed on the story a student tells and how both their in-school and out-of-school experiences have guided their path to medicine. The personal statement and essay responses will be key under this new model. The best applicants will be the ones who show critical reflection on their lives and demonstrate that their choices to partake in certain activities is a result of intentionality, curiosity, and a genuine desire to serve others. These experiences require time, resilience, and a great deal of thought and would tell us far more about a future physician than a flawless transcript assembled through artificial means.
Let’s say Student A worked at various restaurants and served countless patrons doing so. He writes a strong personal statement showing how his experiences improved his ability to communicate with diverse clients and work in a team. His commitment to serving others and ensuring they had a satisfactory evening combined with his scholarly interests towards the sciences blossomed into a desire to pursue medicine. However, he could only get a 3.6 GPA, largely because he was working so much during the semester. Student B had a perfect 4.0 in college, but he used ChatGPT to do so. Student B was a part of a few generic pre-med clubs and cannot speak as much to his experiences in his essays or interview. Under our current system, Student A is likely to be viewed as less than or the same as Student B, which is the crux of the problem.
In addition we should also encourage the return of in-person interviews, which can help admissions committees further understand these softer components of the application. In an age where AI will always have more knowledge than any human, demonstrating how you think and your ability to communicate will likely become a physician’s most vital skill. Multiple-mini interviews (MMIs) and other problem-solving interview formats have become more commonplace in medical school admissions and may serve as another useful tool.
Higher education is at an inflection point right now. We must acknowledge that AI technology will be a part of our lives going forward and rethink what we go to college for in the first place. The same reckoning must also be done by our nation’s medical schools. I see this as an opportunity to break free from our rigid, outdated pre-med system and create a new definition of what a competitive applicant is. And it is not the person who can plug-and-chug organic chemistry questions into ChatGPT and score a 100 on their take-home quiz - it is the person who can effectively tell me who they are and why they want to serve others.
What I’m Reading, Watching, and Listening To:
On Writing About Science in 2025 - Rose Hayes
Using Anonymous Social Media in Modern Medical Education - John Waters
Vinay Prasad Speaks to CBER Staff: Which Prasad Showed Up? - Jeremy Faust