AI saved clinicians the equivalent of 16 working days last year. That number made headlines. It traveled through conference decks, health system newsletters, and LinkedIn feeds as proof that the future of medicine had arrived.

What traveled a little more quietly was the number right next to it: 70% of those same clinicians received little to no training on how to use the tools that saved them that time.

That second number is the one we need to talk about.

The Misconception at the Center of This

There is a narrative gaining ground in healthcare right now — comfortable, well-intentioned, and I believe genuinely dangerous — that AI is something that carries you. That it closes the gap between where your training ends and where the patient needs you to be.

It doesn't. It was never designed to.

A tool is only as sound as the judgment directing it. And judgment — real clinical judgment, built over years of training, earned through experience, sharpened by the cases that didn't go cleanly — is not something any model can generate on your behalf. When a clinician leans on AI to compensate for what they haven't learned, they are not more capable. They are more exposed. And so is their patient.

This is not a critique of the technology. It's a critique of how we are choosing to introduce it.

What AI Is Actually For

At its best, AI is an extension of a trained mind. It handles what doesn't require clinical expertise so that expertise can go where it's actually needed. It surfaces patterns across data that no human could process at the same speed. It reduces the administrative friction that has quietly been consuming the hours clinicians trained for years to use.

But extension requires something to extend from.

A scalpel in an untrained hand is not a surgical instrument. It's a risk. The same logic applies here, and the stakes are identical — because in both cases, the person on the other side of the tool is a patient.

The 64% Problem

A separate finding from the same research deserves its own moment: 64% of clinicians are reaching for personal AI tools (consumer applications, off-platform solutions) when the tools their institution provides don't meet their needs.

I understand that instinct completely. Organizations have not kept up. The training infrastructure isn't there. The approved tools often lag behind what's available. And so clinicians adapt, the way they always have, with whatever is in front of them.

What worries me is when adapting starts to look like substituting. When the tool becomes the answer rather than the instrument. When a clinician stops asking whether their judgment is sound and starts asking whether the output looks right.

Those are different questions. The first one is medicine. The second one is a delegation of responsibility that no AI system is equipped to accept.

Who Needs to Lead This Conversation

The training gap is real. Institutions are behind. Policy hasn't caught up. All of that is true, and all of it needs to be addressed at a systems level.

But in the absence of that infrastructure, the question becomes: who is qualified to bridge it?

Not the technology companies. They can build the tools but they cannot teach clinical application from the inside of a practice they've never been part of. Not the administrators. Not the consultants who have read the literature but have never had to make a decision in real time with a real patient depending on the outcome.

This conversation needs to be led by clinicians who are still in the room. Still seeing patients. Still carrying the weight of these decisions firsthand. Clinicians who understand both what the technology can do and what it absolutely cannot replace.

Why We Built Mentor-IA

That is exactly why my colleague Dr. Javier Mendoza and I founded Mentor-IA.

Between us, we bring decades of combined clinical experience — Rafael as a general, trauma, and robotic surgeon practicing in the United States; Javier as a gastroenterologist with 20 years of international experience across Europe, with a focus on digital health, AI, and medical ethics. We are not observers of this transition. We are living it, in clinical practice, every day.

Mentor-IA exists because we believe that healthcare AI education should be built by clinicians for clinicians — grounded in the real doubts, real complications, and real ethical weight that only people who have practiced medicine actually understand. Not designed from the outside looking in.

Our work spans keynotes, seminars, workshops, and consultations — all oriented toward the same goal: helping clinicians harness AI in a way that elevates their practice, their outcomes, and their patients' experience. Not shortcuts. Not substitutions. Extension of what they already know how to do.

The Bridge

AI is a bridge. A genuinely powerful one, built with extraordinary care by people who understand what medicine needs.

But it was always meant to be one you cross yourself.

The clinician who arrives on the other side having been carried there is not more capable than when they started. They are in unfamiliar territory without a map.

The clinician who crosses it — trained, prepared, judgment intact — arrives somewhere entirely new.

That is the difference Mentor-IA exists to make.

Learn more at mentoria.health

——————————————————————————————————————

Dr. Rafael Grossmann is a trauma surgeon, digital health innovator, and global keynote speaker focused on the intersection of technology and human-centered medicine. He speaks on AI in healthcare, physician burnout, and the future of patient care.