The 3 most promising uses for GenAI in healthcare

Imagine a healthcare system where patients receive world-class medical expertise without stepping into a doctor’s office or even talking with a clinician. Where chronic diseases are monitored continuously and controlled effectively from home. Where medical costs are affordable for all and hundreds of thousands of lives are saved each year. Where burnout no longer drives doctors from the profession.

This future isn’t fiction. It’s within reach—if healthcare leaders embrace the accelerating power of generative AI with bold, strategic action.

I had the opportunity to describe this trajectory during a recent expert webinar hosted by GLG. Speaking to a live virtual audience of more than 600 healthcare leaders, clinicians, investors and business professionals, I outlined the three most promising applications of generative AI in American medicine—models that will transform healthcare over the next five to 10 years.

Each is powerful on its own. But taken together, they offer a roadmap to address the most critical challenges in our healthcare system: skyrocketing costs, subpar quality, poor access and rising clinician burnout.

1. Large Language Models (LLMs): AI as a source of medical expertise

The first and most foundational innovation is the large language model—the technology behind tools like OpenAI’s ChatGPT and Google’s Gemini. These models are capable of synthesizing information from across medical literature and distilling it into patient- or doctor-friendly advice.

What makes this a game-changer is not just the breadth of knowledge these tools contain, but their ability to personalize information at scale. Already, studies show that LLMs can match or exceed the diagnostic accuracy of the average physician. But more important than the ways generative AI will assist clinicians is how these large language models will empower patients.

For decades, clinicians have called for more engaged, informed patients. Now, with GenAI, people can better understand their symptoms, ask questions in plain language and become active participants in their own care. It’s a profound shift in the doctor-patient relationship—one that promises to enhance trust, comprehension and outcomes.

2. Complementors: Combining GenAI with human clinicians in real time

The second major innovation involves pairing GenAI with other technologies that amplify its impact—what I call “complementors.”

The most powerful example is telemedicine. Individually, telemedicine and GenAI have limitations. Telemedicine provides access to doctors but can produce excess demand on already busy doctors. GenAI can monitor symptoms continuously, but it can’t prescribe medications or admit patients to hospitals.

Together, however, they overcome the medical limitations of time and distance.

GenAI can serve as a constant source of expertise for patients, analyzing data and flagging potential issues 24/7. And when an issue arises, virtual visits offer patients instant connection with a live clinician—regardless of where the person is or the time of day. This partnership overcomes the constraints that have long held back traditional medical care.

By combining the strengths of both, this complementor model promises to deliver care that is faster, more accessible and more affordable, without sacrificing quality or compassion.

3. Derivatives: Hyper-specialized AI agents trained on real-world medical interactions

The third model builds on LLMs through a process known as knowledge distillation, a technique popularized by the Chinese company DeepSeek AI. In essence, it involves extracting general knowledge from a large foundational model, thereby allowing smaller companies to build generative AI tools for specific use cases.

These “derivative” agents can then be trained on real-world medical interactions—conversations between patients, nurses and physicians in advice centers or chronic disease programs. The result: Highly targeted GenAI assistants capable of offering medical advice or helping patients to better control ongoing conditions like diabetes, hypertension and heart failure with remarkable precision and consistency.

According to CDC estimates, better control of chronic diseases could reduce their complications (including heart attacks, strokes, cancers and kidney failures) by 30-50%. This would save up to 500,000 lives and $1.5 trillion annually in the United States.

Clinicians, patients and AI: a dynamic trio

In the Q&A portion of the webinar, we discussed challenging and complex topics including privacy, security, safety, bias, liability and the potential impact on medical jobs.

I’m optimistic that these tools will not replace doctors. And I’m confident many of the attendees at this event will be harnessing the power of generative AI to transform medical care in the near future—improving the lives of patients and doctors.

I want to thank GLG for the invitation to speak, and for organizing such a thoughtful conversation. Special thanks to Caroline Ashmore, Danielle Stolakis and Blythe Bailey for their expertise and support throughout.

We are at a turning point in American healthcare. The real challenge now isn’t the power of the technology. It’s the willingness of leaders to act boldly, think long-term, and build the future that patients and clinicians deserve.

* * *

Dr. Robert Pearl is the former CEO of The Permanente Medical Group, the nation’s largest physician group. He’s a Forbes contributor, bestselling author, Stanford University professor, and host of two healthcare podcasts. Check out Pearl’s newest book, ChatGPT, MD: How AI-Empowered Patients & Doctors Can Take Back Control of American Medicine with all profits going to Doctors Without Borders.

The post The 3 most promising uses for GenAI in healthcare appeared first on Robert Pearl, MD.