Is artificial intelligence going to replace physicians? I think the most accurate answer that can be given at this point is “maybe.” Either way, vehemently denying the mere possibility serves no one. If the takeover does happen, being proactive (rather than reactive) and carving a meaningful space for the physician workforce in light of AI leaves us better off than sitting on the sidelines as the next chapter begins. And if the takeover doesn’t happen, thinking about how we would have adapted if it had happened can get us closer to understanding what really matters in medicine.
Don’t get me wrong; I’m scared of AI replacing my future profession. But I’m not worried, because fear can sharpen our focus rather than paralyze us—if we’re willing to look past it and into the uncertainty.
We don’t know what an AI medical field might look like. My mind wanders to hallucinations leading to misdiagnosis, patients (unknowingly) misleading chatbots into prescribing the wrong medication, and health equity nightmares. But if we accept, for the sake of argument, that AI really could one day outperform the median physician on all counts, we may need to apply a lens of humility and rethink what counts as good care. The uncomfortable truth I’ve come to face is that if AI can one day make more correct diagnoses, perform procedures with fewer complications, and provide more equitable treatment than we do, patients may be better off with AI than with us. After all, are we practicing medicine for the preservation of our own careers? Or are we practicing for patients?
You may be shaking your head. Most of us deny that some Silicon Valley executive will ever (not in a thousand years!) have the last laugh over physicians, whether because of liability issues, legislative constraints, data collection limitations, and so on. Much ink has been spilled on these objections, so I won’t try to convince you any further. Rather, I want to suggest that there’s some value in at least entertaining how we may have to adapt if an AI takeover does happen at some point in the distant future. That is, what could we do to make sure that even if algorithms win the arms race, we’re not left holding the pager.
I believe that physicians can evolve through an AI takeover. Like any evolution, it will take time. Some areas of medicine might remain in human hands for longer, allowing a gradual shift. How will a shrinking physician workforce adapt, if it’s not to disappear entirely?
Physicians today play a crucial role in the hospital, beyond making diagnoses, prescribing treatments, and writing notes. They are knowledgeable and faithful companions to those traversing the dark terrain of illness. They serve as humanistic interpreters about death, life, and health. Physicians should bring thoughtful attention, compassionate presence, and timely insight to the bedside. That is, physicians should be humanists, in every sense of the word. I suspect that this role—of witness, interpreter, and partner—will remain uniquely human in an increasingly AI world.
Serving patients in this way may not bill as much as even one minute in an operating room, but maybe that’s the baby in the bathwater. What AI might reveal by handling the algorithmic, quantitative aspects of medicine is that the heart of our profession was never about writing a comprehensive note or recalling the most recent guideline for managing a condition. Being there for patients, a human being with another human in need—that’s the soul of medicine. A human voice to explain the intimate details of what’s happening in a patient’s body, a steady human hand to hold at the end: these are the enduring capacities of physicians. In this sense, AI is not the enemy but rather a mirror that forces us to look beyond RVUs and ask ourselves who we are when algorithms can do everything else.
Granted, there are numerous details still in delicate balance. Issues like changing compensation, debt, and workforce size aren’t settled. The real ingenuity, I suspect, is in how we handle these transitions as a profession. For example, how do we prevent debt-laden physicians being launched into a new field with fewer positions available? These problems seem to be aligned with solutions that are helpful not just in some remote AI takeover, but today. We might focus on increasing the accessibility of medical education (e.g., decreasing the cost of medical school) or widening the scope of physician labor (e.g., to biotechnology, consulting, etc.). While I don’t claim to have all the answers for the future of medicine, I’m certain that physicians can survive if we’re willing to look past denial and toward possible futures, even if we don’t like how they seem at first. If we look ahead, as the entrepreneurs and tech companies are clearly willing to, we may be able to carve and protect a niche for ourselves in which we still provide the best for patients.
That is how I learned to stop worrying and love AI. Not through blind trust in algorithms, but in confidence that no matter what comes next, medicine will still need people with the courage to witness suffering and the wisdom to guide others through it. That role, of steadfast compassion, has always belonged to us.
Rajeev Dutta is a medical student.