At the end of the day, the patient will ask- “Doc, should I take the red pill or the blue pill?” Answering this simple question requires a physician’s discernment. I am convinced our ability to discern truth from what is presented as truth will be the moat for the future.

Medicine at its core is metric-driven; we spend our time looking at numbers and making sense of all the medical information thrown at us. In a way, an Internal Medicine (IM) physician’s core job is not too dissimilar to that of a data analyst, parsing through information to identify what’s wrong and then prescribing a solution for the same. That process involves pattern recognition, not exclusively, but it depends heavily on it. That particular step of the process is what AI is so incredibly good at.

Is it perfect? No. Is it better than the average doctor? That’s old news now.

Now, one might assume that since AI can piece together a problem and a differential based on the information it has received, it will very easily prescribe a solution. Where does it look for this solution? Simple Medical Journals.

Given the plethora of science produced daily, AI can look at the data out there and summarize for you what is written in real time. It can help you keep up. But as anyone who dabbles in Evidence-Based Medicine will tell you, so much of the evidence is misleading. AI cannot separate truth from hype. Even LLMs trained on the best edited journals cannot “think” for themselves. They cannot discern for themselves. They will take all information you input as gospel- whether it is text from a research paper or symptoms from a patient.

Information retrieval and pattern recognition are no longer defensible in the modern age. But the ability to look at what’s being told to you, analyze for yourself, and counterargue will become invaluable in a world that outsources more and more of its thinking to machines.

[1]Discernment #1 by Polly Castor