What happens when one of the most transformative technologies in history collides with one of the most complex and high-stakes industries?
At CHM Live on June 30, 2026, journalist and author Katie Hafner sat down with Dr. Robert Wachter, chair of the Department of Medicine at UCSF and author of A Giant Leap: How AI Is Transforming Healthcare and What That Means for Our Future. Together, the husband-and-wife duo explored how generative AI is already changing healthcare, and why Wachter is surprisingly optimistic about what comes next.
The event was sponsored by the Patrick J. McGovern Foundation.
With the rise of AI, is healthcare finally at a turning point? Watch a CHM Live book talk with Robert Wachter, MD, chair of the Department of Medicine at UCSF and one of the nation’s leading voices on healthcare innovation.
From Digital Records to AI
Wachter has spent decades studying the impact of technology on healthcare. His 2015 book, The Digital Doctor, examined medicine’s transition from paper records to electronic health records (EHRs), a change he once believed would transform healthcare.
Instead, he found that while digitization was necessary, it often created new problems. Doctors became overwhelmed by documentation requirements and spent more time looking at screens than at patients.
The arrival of generative AI in late 2022 felt different.
Initially reluctant to write another technology book, Wachter worried that AI was evolving too quickly for any book to stay current. Ultimately, he decided to focus not on specific technologies, but on the larger question of what happens when increasingly capable AI enters a field where lives are at stake.
Unlike earlier forms of AI, large language models can work with the text, conversations, and medical literature that make up much of modern medicine. That capability, he argued, opens the door to changes that were impossible just a few years ago.
The End of the Data Entry Doctor
One of the most visible early successes for AI in healthcare is something many patients have already encountered: AI scribes.
After obtaining patient consent, clinicians can record conversations during appointments. AI systems then generate clinical notes in the format required by electronic health records.
The technology does more than transcribe. It reorganizes conversations, identifies medically relevant information, and creates documentation that would have previously required significant physician effort.
For Wachter, the impact is simple. Doctors spend less time acting as data-entry clerks and more time paying attention to patients.
Beyond note-taking, AI tools are already helping physicians summarize lengthy medical records, draft discharge summaries, and answer clinical questions in real time.
Wachter described regularly using AI-based medical reference tools to access specialist-level knowledge at the bedside, helping him make better-informed decisions while caring for patients.
Better Than Perfect
One of the central themes of the discussion was how society should judge AI in healthcare. Wachter argued that comparing AI to perfection misses the point.
The current healthcare system, he noted, struggles with clinician burnout, long wait times, rising costs, and preventable medical errors. The right question is not whether AI is flawless, but whether it can improve on the status quo.
To illustrate the point, he compared healthcare AI to self-driving cars. Driverless vehicles receive intense scrutiny whenever they make mistakes, yet the overall safety record may still exceed that of human drivers.
The same standard, he suggested, should apply to healthcare.
At some point, AI will make mistakes. The more important question is whether it ultimately helps create a safer, more effective healthcare system.
Trust, Empathy, and the Human Connection
Perhaps the most surprising topic of the evening was empathy.
For years, many physicians assumed that communication and emotional connection would remain uniquely human strengths. Yet studies have shown patients sometimes rate AI-generated responses as more empathetic than those written by physicians.
That finding has forced healthcare professionals to rethink assumptions about what AI can do.
Still, Wachter believes there are limits.
While AI can simulate empathy remarkably well, he doubts most patients are ready for a chatbot to deliver life-changing news or guide them through serious illness without human involvement.
The challenge, he said, is figuring out how to balance the strengths of both.
As AI grows more capable, trust becomes increasingly important. Patients naturally trust people more than machines, yet modern AI often feels like a person in conversation. That humanlike quality makes it harder to know when confidence is warranted, and when skepticism is needed.
Teaching the Next Generation
The conversation also explored how AI may reshape medical education.
Students and residents now have access to tools that can answer clinical questions instantly, creating concerns about “deskilling” or even “never-skilling,” where future clinicians fail to develop expertise because AI is doing too much of the thinking.
Wachter believes AI should become part of medical education, but not a replacement for learning. Physicians still need to know how to frame problems, identify relevant information, and evaluate whether an AI-generated answer makes sense.
As he noted, powerful tools are most useful in the hands of experts.
Will AI Replace Doctors?
Despite the rapid progress of AI, Wachter’s answer was ultimately no.
To explain why, he shared a story from the book’s final chapter. While caring for a young woman with severe autism who had been hospitalized with pneumonia, the diagnosis and treatment were relatively straightforward. AI could likely have handled those tasks.
But medicine involves more than diagnosis.
As the patient recovered, her mother celebrated her return to her “sassy self.” Later, the young woman smiled at her care team and blew kisses.
For Wachter, moments like that highlighted something AI cannot replace: the uniquely human relationships at the center of medicine.
He left the audience with a vision of a future where AI handles many routine tasks, improves access to care, helps clinicians make better decisions, and lowers costs. Yet even in that future, he believes patients facing serious illness will still need human guides capable of judgment, compassion, leadership, and connection.
As powerful as AI may become, he argued, there will still be times when patients need a doctor.
