Getting AI in Healthcare Right: Nicole Junkermann's August 2026 Take on Thoughtful Oversight

The most encouraging news in health is rarely a single breakthrough. Sometimes it is simply a roomful of thoughtful people agreeing to slow down and get something right.

This past week, a new national effort brought clinicians, patients, regulators and technologists into the same conversation, all of them trying to shape sensible rules for artificial intelligence in healthcare. It would be easy to file that under bureaucracy, one more committee in a field that already has more than a few. I see it differently. When a country decides to pause long enough to ask how a powerful tool should be used before it is everywhere, that is not caution for its own sake. It is a sign of seriousness, and it is good news for the people these tools are meant to serve.

Oversight is a feature, not a brake

There is an old assumption that rules and progress pull against each other, that every safeguard is a cost paid by innovation. In healthcare I have never found that to be true. The best oversight does not stand in the way of good technology. It clears a path for it. When there is a shared, honest understanding of what safe enough actually means, the people building responsibly can move with confidence rather than second-guessing themselves at every turn. Clarity is a gift to the careful. It is only the reckless who have anything to fear from a sensible rule.

Trust is the quiet currency of all of this. A tool can be brilliant on paper and still fail if the person it is meant to help does not believe in it. Patients accept new things when they can see that someone has thought hard about the risks on their behalf. Clinicians lean on a system when they understand where its judgement ends and their own begins. Good rules are how that trust gets built in the open, and once it exists, adoption follows far more smoothly than it ever does when trust is simply assumed. The plain-language glossary kept here exists in the same spirit, to keep the words honest.

Everyone in the room

What struck me most about the week was who was invited. Not only the people who build these tools, but the people who use them, live with them and answer for them. That mix matters more than any single clause that comes out of it. A clinician knows where a busy day actually breaks down. A patient knows how it feels to be on the receiving end of a decision. A technologist knows what the tools can and cannot yet do. And those charged with drawing the lines know how to turn all of that into something durable. Each holds a piece of the picture that the others simply cannot see.

When those perspectives meet early, before a technology has hardened into a hundred finished products, the result tends to be tools that fit real life rather than tools that ask real life to bend around them. That is the difference between something imposed and something designed with care, a theme that runs through the human stories gathered on this site.

What careful actually looks like

It helps to be concrete, because careful is not a mood. It is a short list of habits that any good effort in this area tends to share:

• Ask what could go wrong. The most useful question in the room is rarely how well the tool performs at its best, but how it fails, and who is affected when it does.
• Keep a person accountable. A machine can suggest, sort and flag, but a human being should remain answerable for what happens to another human being.
• Explain the decision. A recommendation that cannot be put in plain terms is hard to trust and harder to correct.
• Test gently, on real life. The proof of a tool is not the demonstration but the ordinary week, measured honestly and adjusted without ego.

None of these are exotic. They are the same instincts any thoughtful person brings to a decision that matters. Written down and agreed on together, they become a foundation everyone can build on, which is a great deal more useful than each team quietly working it out alone.

It is a good moment to be optimistic about this, and not because any one rule will be perfect. It will not be, and it will be revised, as good rules always are. The reason for hope is the posture behind the week: a willingness to bring the right people together and to get the foundations right before the building goes up. Artificial intelligence in healthcare is going to be one of the most consequential technologies of our lifetimes, and the places that greet its arrival with this kind of calm, inclusive seriousness are the ones whose patients will benefit the most.

I have written before, in these notes gathered under the name Nicole Junkermann, that the most human thing technology can do in medicine is to earn its place slowly and keep people firmly in charge. A week spent setting thoughtful rules is exactly that idea in practice. For anyone who wants to follow the thread further, the wider blog keeps returning to it, and the wellbeing tech written about here comes back, again and again, to the same quiet belief: that care done well is care that keeps the person at its centre.

← Back to the blog