After the Aura: What Doctors Are Actually For

Earlier this year, in an essay for JAMA, I argued that artificial intelligence is dissolving something old and important in medicine — what I called the physician's aura. I knew the argument would prompt disagreement. I did not expect quite this many podcast invitations.

The most recent of those, with Bryan Pilkington on the Health Ethics Podcast, gave me forty minutes to take the argument further than the journal page allowed. If you would rather hear the conversation than read me write about it, the full episode is here. If you want the original JAMA piece, it's here. This post is for the reader who wants the short version of what I came away thinking after the conversation.

The argument in JAMA is straightforward enough. For most of medicine's history, the physician's authority drew on a kind of mystique — we knew things patients could not easily check. AI dissolves the mystique. A patient now arrives at the exam room having already consulted an algorithm that knows more medicine than I do. The aura, in any meaningful sense, is gone.

The instinctive response — among my colleagues, and in much of the bioethics literature — is to mourn the aura and ask how to get it back. I think that is the wrong move.

Here is what I argued on the podcast, and what I want to set down in writing.

The aura was never the value. When the stethoscope arrived in the early nineteenth century, the public worried that listening through a tube would replace doctors' judgment. When the X-ray arrived, the same worry. The MRI. The electronic health record. Each time the technology was greeted as the end of the physician. Each time the role survived — but it changed. The doctor stopped being the keeper of certain facts and became something else. We are in another one of those moments, and AI is the technology that will tell us what comes next.

What doctors actually do is sit with uncertainty. A patient with a phone now has access to more medical information than any physician could memorize. So information was never what we were selling. What a doctor does, what an algorithm cannot do, is help a frightened person hold an uncertain decision steady enough to make. We absorb their fear. We tell them what the numbers mean for them, in their life, with their values. That part of the job is not going anywhere. If anything, AI will make it more central — because once the algorithm has produced its probability, someone still has to sit in the room with the person who received it.

The next role is interpreter and steward. I do not think the doctor of 2035 is a smaller figure than the doctor of 1985. I think she is harder to be. She has to know enough medicine to evaluate the AI's output. She has to know enough about the patient to translate that output into something usable. And she has to be willing to disagree with the algorithm in public, in front of the patient, when her judgment says the algorithm is wrong. That is a harder role than the old one, not an easier one.

If you are a patient reading this — and most of the people who reach this blog are, in one way or another, patients — here is the practical thing I would say. Your doctor is going to walk into the room in the next few years having already seen what an AI thinks about your case. Push back. Ask what the algorithm got right. Ask what it might be missing about you. A good doctor will welcome the question. That is how the relationship survives the technology — not by pretending the technology is not there, but by working out together what to do with what it says.

The aura is gone. What replaces it, I think, is something better: a more honest conversation between two people, with a powerful tool on the table between them.

Listen: Health Ethics Podcast — Artificial Intelligence & the Social Role of Medicine Redefined (40 min, hosted by Dr. Bryan Pilkington).

Read: The Lost Aura of the Physician in the Age of Artificial Intelligence — JAMA, April 2026.