Recently, a Harvard-led study published in Science reported that an AI model outperformed two attending physicians in diagnosing real emergency room cases at the initial triage stage. In that study, the AI produced the exact or very close diagnosis in 67% of cases, compared with 55% and 50% for the two physicians. The question now becomes: if AI continues to improve in technical tasks such as triage and diagnosis, will it begin to replace physicians in some parts of medical practice?
There’s a lot to unpackage here and there isn’t a clear simple answer, but I’ve made a lot of observations over the decades about medicine and medical training that are concerning me. Now in 2026 we find ourselves in a unique position where medicine is right now and where it’s going and I’d like to share some of these thoughts with you.
Why Physicians Are Being Treated Like Technicians in Modern Medicine
Back in the ‘70’s when I was trying to get into medical school, one of my college professors stated that physicians were no more than technicians. Unfortunately, today, many physicians do act as technicians, viewing a patient’s medical issues just a problem rather than a human being that is suffering. Patients have complained to me that they were looked at as if they were nothing more than a problem to be addressed rather than as a human being.
This approach is troubling to me, because when a patient has a level of trust in their physician, they are more likely to open up to the physician. This is extremely important when you are dealing with patients with multiple medical problems along with life’s stressors that may have a negative impact on them getting better.
For years I’ve employed a technique that I call, ‘listen with your eyes and see with your ears’. Body language can tell you so much about how comfortable patients are in telling you their story. Their tone can speak volumes. Trust is key in a doctor-patient relationship and totally observing your patient during the history taking of the exam can help dramatically in getting your patient better.
How Eroding Patient Trust and AI Diagnostics Are Reshaping the Doctor-Patient Relationship
In addition to AI outperforming doctors in the ER in 2026, there are other things that concern me about my profession. Trust in the medical establishment has eroded. According to a recent Gallup poll, the percentage of Americans rating medical doctors highly for honesty and ethics fell to 53% in 2024, down sharply from pandemic-era highs. That decline matters, because once trust is weakened, patients are less likely to believe recommendations, follow treatment plans, or feel confident that the system is acting in their best interest. The mismanagement of the Covid-19 pandemic did not help. Another factor is the growing public discussion about some patients feeling dismissed, minimized, or invalidated when their symptoms are not fully explored.
There is good evidence that this relationship matters. Recent reviews have found that better doctor-patient communication is associated with stronger medication adherence, better blood pressure control, and improved patient satisfaction. Other research on trust in clinical practice has shown that when patients trust their physicians, they are more willing to disclose important information, seek care earlier, and stay engaged in treatment. In other words, compassion and communication are not soft extras. They are part of what makes medical care effective.
These trends need to be reversed, and one way to do that is to bring compassion back to medicine. AI may continue to improve and may even surpass physicians in certain technical tasks, but healing has never been only about data, speed, or accuracy. Patients need to feel heard, respected, and understood. They need physicians who can recognize suffering, build trust, and respond with both knowledge and humanity. If medicine is to move forward without losing its soul, then compassion must once again be seen not as an extra, but as an essential part of good medical care.
