AI could transform healthcare for the better — but this doctor says there’s one thing it should never do

I would draw the line at consequential clinical decisions where human judgment, context, and values matter. AI shouldn’t independently decide whether someone receives a life-altering treatment, whether care should be withdrawn, or make another important decision about a patient’s health. Those decisions require a clinician who can understand not just the data but also the whole person sitting in front of them.

AI can be extraordinarily valuable in doing what machines do well, such as analyzing enormous amounts of data, recognizing patterns, reducing administrative work, and helping clinicians identify things they might otherwise miss. I’m comfortable with AI informing a decision, but I’m not comfortable with it owning the decision.

As an emergency physician, I think about it this way: I want AI to make my tasks more efficient. I don't want it to make me less empowered to care for the patient in front of me.

Transparency is fundamental to trust in medicine. If AI is meaningfully influencing a diagnosis, treatment recommendation, or other clinical decision, I believe patients must know that.

However, we also have to be thoughtful about what we mean by "AI is involved." AI is increasingly going to operate in the background of healthcare, helping with documentation, workflow, and information synthesis. We don't necessarily need to overwhelm patients with every technology operating behind the scenes.

The distinction I would make is whether AI is involved in influencing the clinical decision that affects the patient. When it is, or when it is used to help inform that decision, patients deserve transparency. Trust is one of the most important currencies we have in healthcare, and adopting new technology can’t come at its expense.

Accountability has to be shared because responsibility for deploying AI is shared. Physicians have a responsibility to exercise clinical judgment and not blindly accept AI’s recommendation. Healthcare organizations have a responsibility to ensure the technologies they introduce are safe, appropriately governed, and continuously monitored. Lastly, technology companies have a responsibility to be transparent about how their products work, where their limitations are, and how they perform across different patient populations.

One of the greatest opportunities for AI is actually one of the simplest - giving physicians more time to be physicians. So much of a physician’s day is consumed by documentation, administrative tasks, and navigating systems that have very little to do with why most of us went into medicine in the first place. If AI can take some of that burden off our shoulders by helping with documentation, synthesizing information, and reducing repetitive tasks, it gives physicians more time and mental capacity to focus on patients.

That matters for the patient, but it also matters for the physician. There is tremendous fulfillment in sitting across from a patient, listening to them, and knowing you helped someone on what may be one of the hardest days of their life. The more technology can give us back the time to do that, the more it can help physicians and our patients.

Where human judgment will always remain essential is in understanding the person behind the data. Patients don't arrive as perfect textbook cases. There are competing risks, incomplete information, personal circumstances, and difficult choices. AI can help inform those decisions, but judgment, empathy, and human connection are what allow a physician to determine what is right for the individual patient in front of them.

I think AI will become incredibly good at helping us diagnose certain conditions, and there may be specific, narrowly defined tasks where it can recognize patterns faster or more consistently than an individual physician. However, I don't think the goal should be to determine whether AI is "better" than a physician.

AI is a tool, and like any tool in medicine, its value comes from how we use it. It can surface patterns, interpret vast amounts of data, and give physicians another source of insight as we evaluate a patient. But physicians can't outsource their judgment to it. We still must question the output, understand the context, and ultimately make the clinical decision. The future I find most compelling isn't AI replacing a physician's expertise. It's a physician leveraging AI to become even more efficient at their work.

I’d tell patients to use AI to become a more informed participant in their healthcare, not to replace their healthcare professional. AI can be a great tool for helping people understand medical terminology, organize questions before an appointment, make sense of information they've been given, or become more knowledgeable about their health.

So, bring the information you’ve found with AI to your doctor. Ask what alternatives exist, what information may be missing, and what their clinical perspective is. I think one of AI's greatest opportunities is not to replace the physician-patient relationship, but to help patients participate in that relationship with more information and confidence.

The most exciting possibility is that the hospital of the future may be much less defined by the hospital itself. Today, healthcare too often requires patients to navigate the system on the system's terms. If AI fulfills its promise, we can begin to reverse that. Technology could help us identify what patients need sooner, connect them to the right level of care, extend clinical expertise beyond traditional hospital walls, and make care more accessible and continuous.

Inside the hospital, I hope AI becomes almost invisible. I can see it soon becoming the default way to handle documentation, make sense of information, anticipate operational needs, and remove friction from physicians' day-to-day tasks. That means physicians and nurses spend less time serving the machinery of healthcare and more time with patients.

The paradox is that the best version of an AI-enabled healthcare system may actually feel more human, not less. And not through holograms or other digitally assisted means. Used ethically and intentionally, AI will give clinicians back the time and capacity to listen, think, connect, and care. If AI can help us accomplish that, that's the future worth building.