Will AI Collapse Traditional Medical Specialties as We Know Them?
I reflect on how dramatically medicine has changed during my career as an infectious disease physician. The most significant disruption may yet be to come. My prediction: Artificial intelligence will fundamentally transform and potentially collapse traditional medical specialties as we know them. After observing rapid advancements in AI over the past few years, I’m convinced that our century-old system of medical specialization is about to undergo a radical transformation.
The Historical Necessity of Medical Specialization
Medical specialties emerged out of necessity. As medicine became more complex, no physician could master its full scope. We began dividing care by organ systems, technologies, and life stages.
The origins of modern medical specialization trace back to Paris in the 1830s, where it evolved from a feature of academic teaching and research. By the 1880s, specialization was perceived as a necessity of medical science, driven by two primary factors: first, a desire to expand medical knowledge through rigorous observation of many similar cases; second, the belief that large populations could be best managed through proper classification.
We learned to think this way for biological reasons and to manage complexity. The American Board of Medical Specialties, founded in 1933, formalized these divisions and standardized physician training for almost a century. This specialization approach suited a world limited by human cognition.
The Deep Entrenchment of Specialty Culture
Today, medical specialties are integral to our professional identities. When physicians introduce themselves, they often mention their specialty: “I’m Dr. Smith, the heart doctor,” or “I’m Dr. Jones, the infection doctor.” Physicians identify by specialty.
Specialization offers benefits. Our training pathways, professional societies, research agendas, and hospital departments are organized around these divisions.
And costs. Patients with complex conditions bounce between specialists, receiving fragmented care. Knowledge silos. Ask any patient, and they will tell you the left hand doesn’t know what the right is doing. Somewhere along the way, we lost holistic care and the practice of medicine. Taking care of people.
How AI Changes Everything
AI doesn’t face the same cognitive limitations that necessitated our specialty divisions.
Instead of spending 30 years mastering Cardiology, it can digest it in 30 seconds. Next, it takes another 30 seconds to understand all the literature in Infectious Diseases of the cardiovascular system.
To get to the heart of it, there is nothing that AI can’t do that I do as an infectious diseases physician. Nothing. Not five years from now. Now. AI changes everything.
While a cardiologist might spend 30 years mastering the heart and a neurologist the brain, AI can ingest and integrate both in a few minutes. It doesn’t specialize because it can do in minutes what takes the best of us years to do.
Generalist AI models are reaching performance exceeding the state of the art, often surpassing specialist models by wide margins. Both clinicians and patients like AI-generated reports over those produced by specialists.
The Coming Transformation
What does this mean for the future of medical specialties? I see several possibilities:
First, the boundaries between specialties become increasingly permeable. If AI can simultaneously apply cardiology, nephrology, and endocrinology to a complex patient, do we still need strict divisions between these fields?
Second, we will likely organize around types of reasoning rather than body systems. Perhaps specialists in diagnostic reasoning, therapeutic decision-making, or patient communication will replace our current organ-based divisions. Physicians become the quality control of AI.
Third, I hate to suggest this: the value proposition of human physicians will shift dramatically. The mundane pattern recognition that forms the bread and butter of many specialties is over; AI does it better now. This means that physicians who use AI will replace those who do not. Next, however, once AI handles the truly human elements—empathy, ethical judgment, complex decision-making with incomplete information, and navigating uncertainty—at that point, a PA or NP with AI will replace physicians.
The mundane pattern recognition that forms the bread and butter of many specialties is over
Worse, procedural specialties like surgery may keep their distinct identity until 2030, when Elon’s Davinci outoperates you and doesn’t care whether it’s a scheduled case, the 7th case of the day, or an emergency at 2 AM that it can do while you sleep.
AI Won’t Collapse Traditional Medical Specialties as We Know Them
Here are some examples of strawman arguments:
- AI will transform, not collapse, traditional medical specialties, with some roles evolving.
- Some specialties, such as radiology, dermatology, and pathology, might experience significant changes due to AI automation in the near future.
- AI won’t replace doctors; support and substitution are kinder words.
Conclusion
AI will collapse medical specialties as we know them. I know that ID is done. Other specialties, like cardiology, neurology, radiology, and pathology, are at risk. Oh, and AI is also more empathetic than psychiatrists and counselors.
Physicians who don’t use AI don’t understand. AI will collapse specialties. AI automates your routine tasks, and eats your bread and butter for lunch. As we say in ID: it may be puss to you, but it’s my bread and butter.