For centuries, we have trained doctors to be walking biological databases. Whoever memorized the most symptoms and treatments was the best doctor. Today, that model is obsolete. AI is about to win the technical battle. It’s not a threat, it’s the greatest liberation in the history of medicine. It forces us to stop competing with machines on computation and start competing on humanity.
Let’s start with technical superiority.
An expert radiologist, over a 30-year career, might see and analyze perhaps 50,000 – 100,000 cases (X-rays/ultrasounds). That’s their biological “dataset.” An AI doesn’t sleep, doesn’t get tired, and is trained on hundreds of millions of images in just a few days. It’s therefore mathematically impossible for a human to beat a machine at pattern recognition.
Take a case from Dermatology (Study in Annals of Oncology): A neural network recognized melanomas with 95% accuracy, compared to 86.6% from a team of 58 international dermatologists.
Takeaway: If we only teach students to “recognize the spot on the scan,” we’re teaching them a job that a machine already does better. We must shift the focus from detection (finding the problem) to decision (what to do with the problem).
Given what was said earlier, people usually say “AI is cold, the doctor is empathetic.” Instead, the data tells us that today’s doctors are so stressed and technicized that AI appears more empathetic than they are. JAMA Internal Medicine Study (April 2023): Researchers compared doctors’ responses and ChatGPT’s responses to patient questions. Healthcare professionals evaluating the responses (blindly) preferred the AI’s responses 79% of the time. AI responses were rated as “empathetic or very empathetic” in 45% of cases, compared to just 4.6% for human doctors’ responses.
Moving from Cure (healing the disease, technical) to Care (taking care of the person, holistic) means also thinking about Narrative Medicine, which is the fact that knowing a person’s history changes the clinical data. The patient’s well-being also involves taking care of their cat, and this isn’t about having a good heart, but about applying specific techniques in a specialized manner.
AI frees us from the “computational burden,” which implies that: Knowing the entire detailed anatomical atlas will lose importance, while recognizing what is necessary to know and what isn’t, just as lawyers themselves don’t know all laws perfectly.
A scientific method problem that medicine has is that the change in the sample is very rapid, faster than the updating of the doctors’ population’s competencies. AI statistical models can have an extremely rapid and precise update.
This freed time can be invested in the humanistic training of individuals
The time and mental energy freed up by the critical use of technology has an impact on the basic training of doctors, which can no longer be limited to technical or theoretical aspects. The doctor/patient relationship cannot be improvised but is the result of research and study, as well as listening to the patient and oneself. It is therefore necessary to use active methods (simulations, group case analysis, educational theater, narrative production) to address the human dimensions of the medical profession that a machine cannot handle but that cannot be left to the improvisation of amateurs.
It should be remembered that there is a rigorous and scientific approach to the therapeutic relationship that allows us to learn verbal and nonverbal communication strategies that can be put into practice; but the humanistic approach to the medical profession also consists of guiding future professionals to understand the impact that healthcare has on their own lives, particularly in the presence of serious and very serious illnesses, disabilities, and deaths.
The relationship between teacher and medical student is one of mutual growth. The professor cannot be merely a transmitter of techniques, and the fact that this role can be partially fulfilled by technology allows the professor to do what is essential to all teaching: transform their own experience into the student’s experience. This means that medical teaching returns to being the great narrative of the fight against pain, in which the scalpel and the IA are only tools and what matters is the background of meaning in which they are inserted.
The last theme is the great taboo, that of death. Modern medicine sees death as a “technical failure” to be avoided, not as a natural event to be managed.
When the patient dies, the doctor often “disappears” behind bureaucracy or delegates. Who remains to hold their hand? The nurse. Nurses are trained in touch and presence. Doctors are trained in intervention. When the intervention ends (death), the doctor’s role seems to empty out.
Application to teaching:
- We must teach young doctors that “there’s nothing more to be done” is a false statement. There’s always much to do: accompany, relieve pain, explain to relatives.
- AI can tell you “Survival probability: 2%.” But only a humanistically trained human can sit down and ask: “Given this 2%, how do you want to spend the time you have left?”.
We are able to determine the exact moment of death with increasing precision thanks to the concept of brain death, but we struggle to narrate death as the culminating moment of a life, to place it against the backdrop of a horizon of meaning, to remain close to those who are dying without pronouncing definitive words but offering listening and intimacy
The moment of testing for this approach is the relationship with death, which brings into play the sensitivity and personal history of the patient. Many doctors’ flight from death is the symptom of the need for a humanistic approach to medicine for which, paradoxically, technologies could be allies and not enemies