AI has become a permanent fixture in the U.S. healthcare system — and it may be closing holes in patient education and engagement.
Twenty-nine percent of adults using AI for health information on a monthly basis according to a recent KFF survey. While some physicians have expressed a general sense of cautious optimism about AI’s potential in the healthcare space, others see gaps between development and the reality of implementation, while others see it as a threat to patient trust.
Leonid Pravoverov, MD, a nephrology specialist at Oakland (Calif.) Medical Center recently joined Becker’s to discuss his thoughts on how AI may actually be a useful tool for connecting with patients and managing medical misinformation while keeping humans at the center of care.
Editor’s note: Responses have been lightly edited for clarity and length.
Question: Many of your patients now arrive with information generated from AI tools, and you said initially you don’t view that as a negative development. I’m curious what your conversations with patients have looked like, and how you strike the balance between meeting them where they’re at with the information, and then also correcting the record sometimes as needed.
Dr. Leonid Pravoverov: My experience is that when patients come with printouts or information they obtained from ChatGPT or another online AI tool, I very rarely see it as presumptuous. I don’t assume they view it as the absolute source of truth. To me, it’s more of a question and a conversation starter.
I actually think it’s great that patients are researching their health and looking for opportunities and options to better manage their care. They’re trying to engage in self-care, and that’s wonderful. I see it as my responsibility to explain how that information aligns with the currently accepted recommendations within the broader healthcare community and within my specialty.
As a nephrologist, my response is usually along the lines of, “That’s an interesting perspective, but this is what the American Society of Nephrology or National guidelines currently recommend.” So, to me, it’s a conversation starter. I don’t see it as an argument, but rather as an opportunity to strengthen patient engagement.
Question: Do you feel like it’s affected patient trust at all, or do you think that that’s kind of remained the same?
LP: If a patient fundamentally mistrusts the healthcare system, there is very little any of us can do in a single visit. Trust is a very personal thing.
What probably matters more is how the physician responds when confronted with questions or challenged by additional information. Some physicians may perceive that as a threat. I personally don’t. I never position myself as someone who knows everything. Just like my patients, I am constantly learning. I use professional resources, databases, and medical literature every day to stay current with evolving standards of care.
To me, this is a wonderful opportunity to involve patients in their own care. We’re constantly talking about patient education, and this is one form of it. The patient isn’t simply waiting for someone to educate them — they’re educating themselves using available resources. What could be better than that?
Question: What have been some of the most impactful uses of AI in your own personal practice?
LP: Ambient scribe technology has probably had the most profound impact. AI-assisted documentation is truly a life changer for practicing physicians. It gives us more time to focus on patients rather than documentation.
I’ve been using different ambient AI platforms for about two years through various pilots. Today it’s mostly about fine-tuning templates, preferences and workflows. The technology is still evolving, but it’s already incredibly valuable. I absolutely love it.
Question: It seems like there’s a lot of flashy things with AI, and then a lot of practical things that give physicians more time with their patients. Where do you see that balance being?
LP: I think there’s still a misconception about the role AI should play in medicine. This is my personal observation, but the early stages of AI and machine learning focused primarily on risk prediction models. Today, I think we’re moving beyond prediction toward supporting decision-making and care delivery.
Many companies make impressive claims that their AI will improve physician workflow or optimize patient care. But when you ask exactly what they’re offering, the answer is usually something like: “Give us access to your medical records, and we’ll generate risk scores or recommendations.”
Personally, I think we’ve largely solved the problem of identifying high-risk patients. Risk prediction is yesterday’s task. The real challenge now is connecting physicians with those identified patients and ensuring that guideline-directed therapies are actually delivered.
Equally important is connecting providers across different specialties and different care settings. That’s where I believe AI is heading — not simply identifying risk, but enabling multidisciplinary care. That’s the transition I see happening.
Question: Is there anything else related to this, or the conversation in general, where you think there are misconceptions? Or anything else that you feel is important to mention related to this topic?
There are so many technologies now being grouped under the broad term “AI.” To me, the biggest challenge is figuring out how to incorporate these tools into clinical practice with minimal disruption to existing workflows.
Physicians don’t need more clicks, more alerts, or more intrusive reminders. We need AI that helps multidisciplinary teams work together and allows information to flow seamlessly.
Fragmentation of care remains one of healthcare’s biggest challenges. Even at the Becker’s meeting in April, many presentations focused on optimizing data or information. But when you looked more closely, that optimization usually occurred within a single domain — laboratory services, radiology, diabetes management, or another isolated specialty.
It’s still optimization within one segment of healthcare.
What we rarely hear about are tools that truly connect specialties and care settings. Complex patients don’t experience healthcare one specialty at a time. Their care crosses nephrology, cardiology, primary care, hospitals, skilled nursing facilities, dialysis units, home health, and many others.
To me, the future of AI in healthcare is not another prediction model. It’s developing tools that create true interconnectivity across the healthcare system and support seamless multidisciplinary care for complex patients.
At the Becker’s 32nd Annual Meeting: The Business and Operations of ASCs, taking place October 29-31 in Chicago, ASC leaders, surgeons and healthcare executives will explore strategies to drive growth, enhance operational performance, navigate reimbursement challenges and prepare for the future of ambulatory surgery. Apply for complimentary registration now.