Five In Five Healthcare

5 Questions in 5 Minutes: Dr. Avi Tsur on AI in Healthcare

Skip to Main Menu

Skip to Main Content

Skip to Footer

Back to Blog

-->

Back to Blog

Can AI predict the unpredictable, transform medical diagnosis, and still honor the art of healing? Hear from Dr. Avi Tsur, Director of the High-Risk Pregnancy Unit at Sheba Medical Center, as he explores how AI is reshaping the future of medicine. In this rapid-fire Q&A, Dr. Tsur dives into the breakthroughs, challenges, and real-world impact of artificial intelligence in healthcare — revealing both the promise and the limits of technology in the clinic.

1. How are doctors and healthcare professionals using AI today?

“It’s amazing to see to what extent we already use AI. Take a pregnant patient of mine: we start by providing better prediction at the beginning of pregnancy, and hopefully soon, better precision as well. That means predicting complications we’re concerned about, such as preeclampsia, fetal growth restriction, or preterm birth.

During pregnancy, we use AI several times. For example, we have patients admitted to our department, and we empower them by leveraging AI tools to help them communicate with their electronic health notes. Then, when it’s time to give birth, we again use AI to improve prediction of labor and delivery complications, helping us make better decisions about the mode of delivery — whether vaginal or cesarean.

After delivery, we use AI to predict postpartum complications, and hopefully it won’t be long before we can also leverage it to predict long-term health outcomes related to pregnancy complications.”

2. What medical challenges could AI potentially solve in the future?

“I believe what will really change dramatically is clinical reasoning.

Currently, we expect a good medical student or resident to focus on the two or three things that matter most — key risk factors, complaints, and findings in the physical exam — and then offer a diagnosis and treatment plan. That’s what makes a good, focused doctor.

Moving forward, medicine will evolve to integrate hundreds of risk factors, the nuances of each patient’s complaint, and data gathered from advanced physical exams that use video and other AI-powered tools. From diagnosis to treatment, the entire process will change dramatically.”

3. What major impact will AI have in healthcare?

“In the AI-driven future, AI will take many tasks from us. It will handle operational work, assist in surgery, and support clinical decision-making. But I think it will take time before AI can replace us in empathy, in communicating a patient’s condition and all its nuances.”

4. What bottleneck is holding AI back in healthcare?

“Our systems have several limitations. It’s hard to integrate new technologies into our electronic health records and workflows.

But there’s another issue, more specific to AI: it lacks transparency. It struggles to provide references for the information it generates, which goes directly against how doctors think. A good doctor wants to see the table or data behind a recommendation. Even a less experienced one still wants to see a textbook or professional-society reference.

AI just writes something. And when you ask, ‘What’s your reference?’ — without naming products — it often can’t answer. In my opinion, reference accuracy is where AI hallucinates the most.”

5. What human skill in healthcare can AI never replace?

“AI will take many things from us, but empathy will take much longer. Being able to tell a patient what you would do in their situation, to share what you’ve done before, or to offer a hug while delivering bad news is our greatest strength.

The parts of our work that are purely technical will eventually be replaced. The parts that are deeply human will take much longer — if ever.”

Discover more

Jun 25, 2026

Token spend isn’t going down. You need more than naive routing to manage it

Labs in Front

-->

Jun 24, 2026

Tipping the scales: Merging weak agents into a state-of-the-art deep researcher

Labs in Front

-->

Jun 4, 2026

First scale, then enrich: How the right execution strategy helped us reach state-of-the-art on SWE-rebench