Can AI Save Healthcare?

When was it that the US Medical establishment stopped providing healthcare? The kindly family doc, making house calls and practicing medicine intuitively, is a relic of another century. Today, what was once a ‘healing art’ has become ‘physiological engineering’.

Yet we are not ‘bionic’ men and women…and we never will be. (AI has already superseded mechanical robotics.) But the healthcare system treats me as if I were a cheap knock-off of the 6 million dollar man.

To be clear, if you have a clearly diagnosed and reasonably serious health condition, you have never been better off. We can do things today with laser knives and designer drugs that were the stuff of science fiction only a generation ago.

But if you’re a poor slob who just doesn’t feel good, you’re sh*t-out-of-luck. Show up at your PCP’s office, at  Urgent Care, or at the ER and you will be put through a battery of generic tests and a lengthy ‘true or false’ intake questionnaire. Your healthcare professional is undoubtedly well intentioned and would like nothing better than to see you leave the building feeling great. But that concern is no longer center stage in medical practice.

For most Americans, healthcare consists of a series of check lists imposed by insurance carriers, government regulators, and our tort system (malpractice). A successful engagement has the patient leaving a doctor’s care with no possible grounds for a lawsuit.

Not that you leave empty handed. Your test and interview scores trigger a series of recommendations that are ‘evidence based’ and reflect ‘best practices’. Your data is analyzed by inflexible algorithms that deliver totally predictable results, but you are receiving the best care that Machine Medicine can offer. 

The current system is well designed to diagnose conditions such as elevated blood pressure or high cholesterol. It is much less suited to discovering the causes of these conditions in a given patient and, if you’re looking for creative ways to treat the patient’s condition, forget about it. And so we treat symptoms and move on.

AI First healthcare is something all together different. Yes, AI runs on algorithms, but it has the flexibility to think outside the box and the capacity to suggest novel diagnoses and treatment strategies. Working with a highly trained professional, AI can bring intuition, the gut, back into the practice of medicine. 

AI is able to build a model of each patient that includes vastly more variables than any human (or any check list) could possibly keep in mind at one time. AI combines virtually limitless memory and instant recall with a full medical school curriculum at its finger tips. The smartest, best trained, and most compassionate human care giver cannot possibly compete.

Plus, medical knowledge is now doubling every 5 years. All the continuing education in the world cannot keep a doctor current, but AI is 100% current all the time. It makes no sense to think that any human doc could compete.

Now to be clear, we are not proposing to remove the MD from the healthcare team; human oversight remains absolutely essential. But in the future, the practice of medicine must be AI First.

Already, AI consumer sites are fielding 50 million health-related inquiries every day. That’s an average of 50 health inquiries per American per year…as compared to the current 20 minute annual check-up. 

Today’s Machine Medicine relies on discrete answers to specific questions but Ars Medica, the art of medicine, requires clinical reasoning capability. To test for that, scientists turned to the New England Journal of Medicine (NEJM) – one of the world’s leading medical publications.  

NEJM publishes Case Records from the Massachusetts General Hospital (MGH), presenting a patient’s care journey in a detailed, narrative format. These cases are among the most diagnostically complex and intellectually demanding in clinical medicine, often requiring multiple specialists and diagnostic tests to reach a definitive diagnosis.

How does AI perform in this arena? Surely here we need the empathy and imagination of a well-trained and highly experienced human agent, right? 

To answer these questions, Microsoft created a series of interactive diagnostic challenges based on 304 recent NEJM case studies. Clinicians, be they AI Bots or human physicians, were able to ask iterative questions and order appropriate tests. As information became available, the clinicians could update their reasoning, gradually leading to a final diagnosis. 

This diagnosis was then compared to the gold-standard outcome published in the NEJM. The Microsoft AI Diagnostic Orchestrator (MAI-DxO) is a system designed to emulate a virtual panel of physicians with diverse diagnostic approaches collaborating to solve cases.  The Orchestrator can integrate diverse data sources. It can turn an LLM into a ‘panel’ of virtual clinicians that can ask follow-up questions, order tests, deliver a diagnosis, and then run a cost check before deciding to proceed. 

MAI-DxO solved 85% of the NEJM benchmark cases. Pretty good! But how does this compare with the recommendations of real life, flesh and blood specialists? For comparison, Microsoft created a panel of 21 practicing physicians from the US and UK, each with 5-20 years of clinical experience. On the same problem sets, these experts achieved a mean accuracy of…wait for it…just 20%. 

Astonishing! AI is 4 times better at diagnosing complex medical conditions than MDs. If Microsoft’s findings are confirmed, traditional medicine is, or should be, dead. The big question now is how long it will take society ‘to stop life support and pronounce’. In the meantime, the surest evidence of malpractice is the practice of medicine without AI support.

Another measure comes from healthcare quality watchdog, Healthgrades. Each year for the past 20, they have monitored patient outcomes at 4,500 US hospitals. According to Healthgrades, if all hospitals delivered care at the quality level of the best 250, more than 200,000 lives would have been saved in 2025: 1 in every 14 in-patient deaths can be attributed to subpar healthcare.

An article by Kristen Brown, published in the New York Times on 1/8/2026 caught the spirit: 

“Amid rising interest in alternative medicine and growing skepticism of physicians, health and wellness companies have begun providing Americans a more direct route to medical testing that does not require a doctor’s visit. 

“In November, the telehealth giant Hims & Hers debuted extensive laboratory testing for customers, just weeks after Oura and Whoop unveiled blood-testing products of their own. The national laboratory network Quest Diagnostics now sells its own consumer-facing tests. LabCorp does, too.

“But perhaps no other company has capitalized as explicitly as Function has on frustrations with the U.S. health system…For $365 a year, the company provides its hundreds of thousands of members with access to more than 160 lab tests.”

Of course, there are downsides to an AI Only approach to healthcare. False positives and unwarranted patient anxiety lead the list. That is why we don’t recommend such an approach. Our AI First model includes the genius of the human physician, interpreting results and presenting them to the patient in context and with specific recommendations. But this can only happen after AI has done its thing.

***

Nurse in Red (1991–1993) by Keith Holmes portrays a nurse in a vibrant red uniform with calm authority and professional poise, conveying her expertise and readiness within a clinical setting. The painting reflects the evolving professional identity of nurses in the late 20th century, highlighting their competence and command in modern healthcare through her composed demeanor and the presence of medical equipment in the background.