Enterprise

UC Health expands AI use for clinical notes, patient calls

System-wide rollout of ChatGPT for medical records and phone agents follows successful pilot programs at UCSD and UCLA facilities.

Omega Editorial · November 20, 2026 · 4 min read

Key takeaways

  • UC Health plans to deploy ChatGPT across all medical center record systems after successful pilots at UCSD and UCLA facilities.
  • AI agents at UCSD Health currently handle colonoscopy prep calls and appointment scheduling, with system-wide phone service expansion planned.
  • Testing with 100 clinicians found AI-generated chart summaries saved physician time in 80% of cases.
  • A medical student regent cited research showing AI documentation tools may impair physicians' ability to recall patient information.
  • UC administrators emphasized thorough testing and monitoring requirements before deploying AI tools across the health system.

The University of California health system is accelerating its adoption of artificial intelligence for clinical workflows, with plans to deploy ChatGPT across all UC medical centers and expand AI-powered phone services that handle routine patient interactions.

Administrators outlined the expansion during a UC Board of Regents meeting this week, describing pilots already underway at UC San Diego Health and UCLA Health that use AI to generate clinical notes, answer patient calls, and prepare summaries from lengthy medical records.

Current deployments across the system

All UC hospitals with medical centers now use AI scribe systems that generate chart notes—summaries of patient visits stored in medical records—according to Suresh Gunasekaran, president and CEO of UCSF Health. The Daily Bruin first reported details of the expansion plans.

At UCSD Health, AI agents currently handle phone calls to help patients prepare for colonoscopies and schedule follow-up appointments, freeing nursing staff for other duties. Dr. Karandeep Singh, chief health AI officer at UCSD Health, said patients have expressed both skepticism and appreciation for the time savings, noting they receive clear instructions through the automated system.

UCLA Health physicians are using AI to document patient interactions, respond to online patient questions, and assist radiologists with diagnostic imaging, according to a June press release from the health system.

ChatGPT integration for medical records

Dr. Smitha Ganeshan, medical director for AI programs at UCSF Health, said her team is working to integrate ChatGPT directly into electronic medical record systems to help clinicians extract relevant information from patient charts before appointments. Testing with 100 clinicians found that AI-generated summaries saved physicians time in 80% of trials.

Once administrators approve the integration, ChatGPT will be added to medical records software across all UC medical centers. If the UCSD Health phone agent pilot proves successful, all UC campuses will receive access to similar AI calling systems, said Dr. David Rubin, executive vice president of UC Health.

Why it matters

The UC system's coordinated AI deployment represents one of the largest academic medical center adoptions of generative AI for clinical workflows. With hundreds of billions of dollars flowing into healthcare AI development quarterly, major health systems face pressure to evaluate these tools while managing concerns about diagnostic accuracy and physician skill retention. The UC approach—piloting tools at individual campuses before system-wide rollout—offers a model for balancing innovation with clinical safety in large healthcare organizations.

Concerns about clinical skills

Student Regent-designate Jordan Roiland, a medical student at UCSF, raised concerns that AI-generated documentation could erode physicians' ability to recall patient interactions. He cited a June 2025 MIT study finding that AI users have more difficulty remembering information when they delegate writing tasks to automated systems.

"We see the most complicated patients in the world at UC," Roiland said. "I'm really worried that this is going to drastically reduce our quality of care and our ability to come up with those very rare diagnoses and hard-to-come-to conclusions when we're pumping through patient visits."

Gunasekaran acknowledged that AI tools often underperform relative to developer claims and said leaders plan thorough testing before implementing any technology across UC campuses.

Dr. Katherine Andriole, associate dean for health AI strategy and innovation at UCLA's Geffen School of Medicine, said UCLA Health monitors AI-generated discharge summaries through large language model review and clinician focus groups. "Adoption and deployment has to be paired with monitoring and governance," she said.

UC medical schools have begun incorporating AI literacy and healthcare AI applications into their curricula, though Andriole noted some students remain hesitant to use the tools.

The Daily Bruin reported these details from the UC Board of Regents meeting held at the UCSF Mission Bay Conference Center.

  • #healthcare ai
  • #clinical documentation
  • #uc health
  • #chatgpt
  • #medical records
  • #patient services

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Enterprise