AI in healthcare: impact on patients and clinicians

‍

What if we could detect serious illnesses before the first symptom ever appears?
What if doctors could finally look up from their screens and connect - truly connect - with their patients?

The question isn't whether AI will change medicine. It's already happening. The real question is: Will it help clinicians care more deeply, or replace human warmth with automation?

In this article, we explore how AI can enhance - not replace - the human side of medicine. From spotting what’s easy to miss, to enabling prevention at scale, to giving doctors more time for care and connection.

We were joined in this conversation by experts: Dr. Anna Drohomirecka, cardiologist and transplant specialist with two decades of clinical experience; Dr. Harvey Castro, emergency physician and prominent voice on AI in healthcare; and Przemek Jaworski, CTO of Shen AI, building tools that bring these ideas to life.

‍

Beyond bias and burnout

Emergency rooms are intense, demanding environments. At 2 AM, when fatigue sets in, even the best ER doctors are still human. No one can perform at 100% around the clock. This is where AI isn’t a rival - it’s an ally. A quiet, tireless, and highly observant assistant. 

It helps clinicians avoid tunnel vision, especially under pressure. As Dr. Harvey Castro points out, doctors are at their best when applying experience - but that experience is only useful once all the possibilities are on the table. AI acts like a real-time checklist, offering diagnostic prompts that surface symptoms or connections a tired mind might overlook.

“It truly helps see things that I cannot see", says Dr. Castro.

‍

Think about something as simple as listening. How often do you listen or observe in a day with the entirety of your attention? When a doctor is stuck locked typing on a computer, they can miss small clues from a patient. Things like a downcast look or a shaky voice, when asked about habits like smoking or drinking, go unnoticed. AI tools like ambient listening free clinicians from constant typing so they can observe, listen, and connect.

Bringing back a human connection that has been missing from medicine.

For this partnership to work, trust must be the foundation. Dr. Castro strongly believes in being open and honest. 

“We should go beyond paperwork that simply hides consent in long forms that people don't read. Instead, imagine a ‘food-like’ label that's there for the patient or a bill of rights at the hospital. The bill would openly state: This is how we're using AI technology. The good, the bad, the ugly of it."

Dr. Drohomirecka agrees, saying transparency is needed both with patients and in how and why medical regulatory boards make rules. Just as a new medicine must go through strict tests, Dr. Drohomirecka emphasizes that experts must verify the bones of the AI they employ. 

“We need to know how the model was built, then on what population it was tested on" she insists. Without this clear information, doctors will find it hard to trust the tools they are being asked to use.

Proactive health: Catching the unseen before it strikes

The big goal of AI in healthcare is to change from fixing what's broken to preventing the break entirely. As Dr. Castro says, “Why wait until someone is diagnosed with diabetes when AI can identify them as pre-diabetic? Why wait for a heart attack when AI can spot a genetic risk years ahead of time?”. Preventive medicine is a huge step forward for global health populations because it hugely lessens the financial burden of illness, meaning we can help more people with fewer resources.

Dr. Anna Drohomirecka paints a clear picture. “Think about the scale of the flow of data from home health monitors – blood pressure measurements or heart rate measurements pour in. For medical staff, going through this massive number of alerts is a never-ending job. AI acts like a smart filter. It precisely picks out only the patients at higher risk that something is going wrong." Helping doctors like us focus on critical, life-saving care for those who need it most.

AI's ability to predict also reaches inside the body. Devices placed inside, like CardioMEMS (a small device implanted in the cardiovascular system), can measure pressures inside the heart. “AI programs within these devices can predict if a patient's heart failure is going to worsen a full week before any clear symptoms appear,” says Dr. Drohomirecka. “This allows doctors to adjust medication and help the patients much earlier."

Even common items like smartphones or watches, when powered by AI, become a warning system for conditions like atrial fibrillation (AF). AI-powered apps are already detecting atrial fibrillation (AF), a condition that often goes unnoticed until it causes complications like stroke. AF is notoriously hard to catch due to its paroxysmal (come-and-go) nature. AI helps capture these fleeting episodes and alert patients and providers before damage is done.

In short: AI sees the "unseen", revealing signals no human could easily detect and helping doctors make smarter, faster decisions. It's a breakthrough in our ability to predict problems early, personalize treatment, and make healthcare not only more effective - but more proactive.

‍

Building trust

For AI to be a valuable part of healthcare, we need to trust it.

Przemek Jaworski, CTO at Shen AI, emphasizes that accuracy depends on the quality of data used to train the model. “It’s not just about quantity,” he explains. “It’s about having representative, high-quality data.” Just as nutrition shapes human health, data shapes AI performance. 

In short, for humans, we are what we eat. For AI, it is the data it digests.

"We first need to make sure that the distribution is correct. That we have the entire spectrum of values... and then that the demographics are inclusive and sound. A balanced data set includes every kind of person – different ethnicities, different skin colors, different lighting conditions, different ages, and many more variables. Jaworski emphasized, “Only then can we start building trust."

Shen AI puts this philosophy into practice with the multimodal sensing approach. This technology combines remote photoplethysmography (rPPG), which detects changes in skin tone from blood flow, with remote ballistocardiography (rBCG), which tracks subtle facial micro-movements from the ejection of blood from the heart. "It is like having two pairs of eyes, seeing the same thing from two different, helpful angles. This dual method greatly improves the quality of measurement," explains Jaworski. Making sure the technology works well for everyone, everywhere.

There is a problem in the broader AI-medical landscape. The complex way AI programs work can sometimes lead to a problem: data drift. Dr. Castro warns that an AI program, even if it works perfectly today, can slowly start changing how it works over time, moving away from its original accuracy. This quiet weakening of reliability is a serious concern that healthcare organizations must address by asking vendors for ongoing maintenance and checks. 

The complex job of adding AI into healthcare needs strong leaders. Dr. Castro suggests a "Chief AI Officer" for hospitals. He argues that this expert is vital to guide AI use, make sure data is correct, and answer important questions about issues like data drift. Still, challenges remain - especially budget constraints and the cost of operating large-scale AI systems.

Education and seamless integration

The path ahead requires a big change in how medical students are taught. Dr. Castro points out that a limited number of schools include AI in their courses. It is a big problem, as AI will be the bread and butter for future doctors. We need a whole shift in the educational system for medical students.

Still, even with AI, the doctor must make the final decisions.

"You have to be a doctor," Dr. Castro insists. Just like questioning a strange lab result, a doctor must check AI's findings, using their experience to tell right from wrong. Experienced doctors, he notes, "Can look at an output from AI and be like, oh, that's wrong," while newer doctors need to learn how to research and verify.

AI that supports - not replaces - people

AI in healthcare isn’t a choice between humans and machines. It’s not about replacing people or simply generating more data. It’s about building a trusted partnership - one that empowers doctors to be more present, accurate, and effective, and gives patients clearer access to their own health.

This future requires openness, strong evidence, and continuous learning. But when done right, it delivers care that’s not just smarter - it’s more human.

Watch the full webinar to hear directly from the experts and explore how AI is reshaping care.

‍