How doctors are using AI to transform clinical care
By Liza Gross, California Health Journalism Fellow
Call it the “Moby Dick” dilemma. The average patient’s chart has expanded exponentially since the early 2000s, when most hospitals adopted electronic health records. “Every chart is now the same size as ‘Moby Dick,’” said Joel Gordon, a family physician and UW Health chief medical information officer, at a Health Journalism 2026 panel in May.
Unlike family doctors, who know their patients’ medical histories, Gordon said, emergency department and urgent care physicians have just minutes to parse a massive EHR file. “It’s cognitive overload.”
Add in the ever-expanding universe of treatments, drugs and risks for a growing proportion of patients with chronic conditions and physicians simply can’t keep up. It’s no wonder they’re increasingly turning to artificial intelligence, or AI, to manage administrative tasks such as insurance authorizations, patient email responses and clinical note summaries.
More than 80% of physicians use AI for work, a 2025 American Medical Association survey found, nearly twice as many as a few years ago, said Michael Suk, immediate past chair of the AMA’s board of trustees.
Panelists at the HJ26 session “How doctors are using AI to transform clinical care.” Photo by Zachary Linhares
Genevieve Melton-Meaux, M.D., Ph.D., speaks at HJ26. Photo by Zachary Linhares
Moderator Karen Blum at the HJ26 session “How doctors are using AI to transform clinical care.” Photo by Zachary Linhares
Most physicians say AI improves their ability to care for patients, Suk said, even as they worry about ethical challenges, data privacy risks and its potential to skew research.
To see if AI tools are working as intended, Gordon’s colleagues developed software to evaluate the effectiveness and efficiency of AI-generated clinical summaries. AI tools can not only save time, they found, but also allow providers to concentrate more on charts, labs and treatments.
Shifting providers’ focus from note-taking to counseling patients is a key benefit of AI tools, said Genevieve Melton-Meaux, a colorectal surgeon and senior associate dean of health informatics and data science at the University of Minnesota.
Yet balancing opportunities to enhance the efficiency and quality of care with the need to protect patients and their privacy remains a challenge, Melton-Meaux said. When your mission is to provide the “most human-centric care” possible, she said, you need policies to ensure that AI applications are safe, ethical and effective.
Beyond its role in health care delivery, Melton-Meaux said, AI is “democratizing knowledge about health” by giving patients access to health information.
At the same time, the panelists agreed, more needs to be done to help patients understand the privacy risks of feeding their personal information into a chatbot.
Toward that end, the AMA created the Center for Digital Health and AI last year to help shape how the technology is used in medicine. The organization also created prompts to help patients get reliable information from chatbots, knowing that people use AI in medical emergencies, Suk said, adding, “I’m guilty of it.”
After Suk’s corgi ate a bar of chocolate, he consulted AI. “Oh my god, my dog ate a bar of chocolate,” Suk told a chatbot.
After asking Suk how much chocolate the corgi ate and weighed, it said, “Don’t rely on me, call your vet.”
Suk arranged a televisit with a vet and was reassured that his dog would be fine. They no longer keep chocolate in the house.
Liza Gross is a California reporter for Inside Climate News based in Kensington, Calif.
