Archimedes said “Eureka,” the ladies at The Knights of Columbus Hall say “Bingo,” and as an ophthalmologist, I say “This is a real eye-opener.” You may or may not recall my exasperated September/October 2025 editorial. The title is self-explanatory: “You have sudden onset of chest pain? The doctor can see you in five months: The travesty of physician unavailability.” Many of you share my pain and concern. In post-publication discussions and correspondence, no physician expressed to me satisfaction with what passes for medical “care” in the United States. Having at times been called a “know-it-all,” in my earlier editorial I disabused that notion as I proffered no solutions. Herein, I rectify that omission.
What is the answer?
Patience, you will have to wait another paragraph or two for my epiphany. In reviewing possible solutions there are many of them out there; all have flaws or impracticalities. To wit:
- Additional medical schools with more students.
- Minting physicians in less training time.
- Funding additional residencies, especially primary care.
- Make present physicians more productive, ideally by greatly truncating the time sink that is present-day electronic medical records.
- Enticing physicians into staying in practice longer (good luck with that).
- Allowing more international medical graduates (IMGs) into U.S. practice.
IMGs have become an indispensable part of U.S. medicine. They admirably often practice in rural or underserved urban areas and in specialties less attractive to U.S. physicians. Left unsaid is the propriety and ethics of taking physicians out of their resident countries that sponsored their education and training. Most of these native IMG countries have greater medical needs and health challenges than our own country.
Lastly, to put it brutally, is to dumb down primary care by turning it over to an eager cadre of often poorly trained and overly confident physician wannabes. The main physician-pretenders are unsupervised nurse practitioners who are nurse educated and to a much lesser extent unsupervised physician assistants that are physician educated. This is also the default direction we are pushed each legislative session. Non-physician scope of practice is expanding faster than the universe. Waits to see a physician are insufferably and dangerously long. Accessibility is the rocket on which enlargement of scope is put into orbit. Dr. Nurse will do tomorrow if Dr. Physician is booked out a half-year or more. An increasing number of people say their doctor is a nurse. Worse yet, they are fine with it.
The AI solution
Thanks for your patience. OK, here it is: the answer. No, it is not so dramatic that you need to sit down before reading. I, along with millions of others, feel the only hope for better, more accessible, less expensive medical care, here, there, and everywhere is (drum roll) artificial intelligence (AI). Maybe you feel the same way; I hope so. I have been cogitating on the U.S. medical crisis incessantly since I care for my 80-year-old wife with severe Parkinson’s disease and cognitive impairment. In my calculus, every health care “fix” fails except medically oriented, patient- and physician-friendly, constantly improving AI. Others have come to the same conclusion.
In 1992, I was visiting my daughter, a student at Stanford University. She took me to the Stanford library, sat down at a computer, and said, “Dad, this is called the World Wide Web and it is going to change how the world works and thinks.” How right she was. For better, for worse, or for both, AI is changing the world exponentially faster than anyone can imagine. If you have not already warmly embraced AI, you are at serious risk of becoming a Luddite.
AI can already outperform most physicians in diagnosis. In several studies, AI did better by itself with diagnosis than when teamed with a physician. Yes, the physician was holding back the AI. Not all AIs, like physicians, are equally talented. At the threshold of 2026, the Einstein of AI is the Microsoft AI Diagnostic Orchestrator (MAI-DxO). It correctly diagnoses 85 percent of cases described in the New England Journal of Medicine. On the other hand, physicians diagnosing the same cases were correct only 25 percent of the time. Wow, 85 to 25; that qualifies as a physician trouncing. AI, circa 2026, exhibits its highest dominance in pathology, radiology, and dermatology. Going forward, AI will come out on top more often in all specialties.
Thanks to AI, we are on our way to the health care promised land: more physicians providing medical care higher in quality and quantity than now, at lower costs, in any venue, be it near or be it remote. Oh, I seem to have forgotten to mention time-appropriate: immediate for true emergencies, same day for urgent/painful conditions, and nonemergent specialty referrals one to two weeks. Health care will be unrecognizably better than today’s “The doctor maybe can see you in maybe six months.”
A vision of the future
To truly appreciate the scale of this paradigm shift, we must visualize the clinical environment that this infusion of medical intelligence will create. The future AI-powered physician’s office will operate as a hyper-efficient, compassionate digital engine, managing rapid, on-demand appointments where a highly advanced AI system performs the initial triage, orders necessary lab and imaging studies (like X-rays), and makes preliminary diagnostic assessments based on exhaustive data analysis. This AI, capable of fluent and empathetic communication in any language and dialect, will explain the diagnosis to the patient with tailored kindness and patience, instantly determining if an urgent transfer to a larger facility is required, yet at every stage, the human element remains paramount, with trained staff instantly available should the patient feel uncomfortable, and a qualified physician reviewing every complex decision, ready to perform invasive procedures and provide the final, human-validated sign-off on the care plan.
(Disclosure: This last paragraph was written by AI and is probably the best paragraph in this editorial. In an example of AI hallucinations most of the references AI cited were fiction. Like children, AI not infrequently makes things up. The prose in the paragraph is exceptional; AI is also swiftly killing creative writing.)
I share the health care grievances of our patients: too expensive, too long of a wait, too little time with the physician, too many errors and human shortcomings (lack of sympathy, uncommunicative, etc.), and too far to drive. Whether you are a consumer or a provider or both, only AI offers the means of getting our sick health care system well.
John C. Hagan III is an ophthalmologist.






![Smart design choices improve patient care outcomes [PODCAST]](https://kevinmd.com/wp-content/uploads/Design-4-190x100.jpg)