Medical school taught me an extraordinary amount about medicine. I learned anatomy, physiology, pathology and pharmacology. I learned how diseases develop, how diagnoses are established and how treatments are chosen. 

Years later, after working as a physician, academic and healthcare executive, I have come to realize that some of the most important things I eventually needed to practice medicine were precisely the things medical school found hardest to teach.

How do you make a decision when the evidence does not provide a single answer? How do you explain uncertainty to a frightened patient without turning uncertainty into fear? How do you recognize when a guideline describes the average patient well but the person sitting in front of you poorly? And how do you accept responsibility for a decision when several reasonable options remain?

These are not peripheral skills. They are part of what transforms medical knowledge into medical practice. This distinction has always existed, but artificial intelligence is making it newly visible. AI can retrieve information, recognize patterns, summarize evidence and increasingly support diagnostic and therapeutic decisions with extraordinary speed. 

As these capabilities improve, an uncomfortable question emerges for medical education: If access to knowledge becomes easier, what should we be teaching future physicians that access to knowledge alone cannot provide?

Taiwan has a particular reason to ask this question, and my interest in how the country approaches it is not incidental. I have visited Taiwan around fifteen times over the past twenty-five years, and those visits have given me a longstanding interest in a society that has repeatedly shown how quickly technological change can become part of everyday life. I write from the perspective of a frequent visitor rather than an insider, but it is precisely this combination of familiarity and distance that has kept me interested in how Taiwan approaches the intersection of technology, healthcare and education.

In June, National Taiwan University disqualified a medical school applicant who allegedly used AI-enabled smart glasses during an entrance examination. The incident understandably raised questions about cheating and academic integrity. But it also points toward a larger question. When technology can increasingly provide answers, educational systems must reconsider what their examinations - and ultimately their curricula - are designed to measure.

Taiwan is already having this conversation more broadly. As the country invests heavily in AI education, educators are debating whether systems built around memorization and correct answers adequately prepare students for a world in which machines can perform many of those tasks extraordinarily well. The emerging emphasis is increasingly on judgment, verification, problem-solving and the ability to apply knowledge rather than merely reproduce it.

Medicine may be where this transition matters most. A medical student must still know medicine. AI does not make anatomy optional, nor does instant access to information eliminate the need to understand physiology, disease or treatment. 

A physician cannot judge the quality of an AI recommendation without possessing sufficient knowledge to recognize when it might be wrong. But knowledge is the beginning of clinical competence, not its completion. Clinical medicine takes place in the distance between what is generally true and what is right for the individual patient. Evidence describes populations; physicians treat people. Guidelines organize accumulated knowledge; patients arrive with combinations of disease, preference, fear and circumstance that no guideline can anticipate completely. AI will become increasingly capable of helping physicians navigate this complexity. It cannot remove the responsibility to decide what should be done for the person sitting in front of them.

That suggests a different way of thinking about medical education in the AI era. The question is not how much traditional knowledge can be removed from the curriculum because a machine can provide it. The more interesting question is what deserves greater attention once knowledge is no longer the scarce resource it once was.

Students need more opportunities to encounter uncertainty before they are expected to manage it independently. They need cases in which several answers are defensible but none is obviously correct. They need to learn how to explain uncertainty, disagree respectfully, recognize the limits of evidence and make decisions whose consequences belong to real people. Most importantly, they need to understand that saying “I don’t know” can sometimes represent the beginning of good clinical judgment rather than the failure of medical expertise.

This is also why I believe Taiwan’s strength in technology gives it an unusual opportunity. A healthcare system operating within one of the world’s most technologically sophisticated societies need not choose between better AI and better physicians. It can ask a more ambitious question: What kind of physician becomes more valuable as machines become more capable?

I suspect the answer will look surprisingly familiar. The physicians who matter most will still be those who know medicine deeply but also know when knowledge is insufficient; who use technology confidently without surrendering judgment to it; and who understand that the final responsibility of medicine is not to produce an answer, but to care for a person. Those are among the most important things I did not learn in medical school.

(This piece reflects the author's opinion, and does not represent the opinion of CommonWealth Magazine.)

CommonWealth Magazine welcomes op-ed submissions. Please send your article proposals to [email protected]

About the author:

Mathias Goyen, M.D., is Professor of Diagnostic Radiology at the University of Hamburg and Chief Medical Officer at GE HealthCare. He writes about medicine, technology, education and the human dimensions of healthcare.

Have you read?

• When Did We Stop Waiting for Answers?
• How Taiwan Can Build Global Brands Without Losing Itself
• Line Logic: Why Trust in Taiwan is Earned Not Announced

Uploaded by Ian Huang