New Study Finds Nearly Two-Thirds of U.S. Hospitals Using Epic Have Adopted Ambient AI—But Disparities Exist

Ambient artificial intelligence tools that capture clinician-patient conversations and generate draft clinical notes are now widely used in U.S. hospitals but unevenly adopted, according to a new nationwide study from researchers at Rollins School of Public Health.

The study, published in the American Journal of Managed Care, looked at 2784 U.S. hospitals using the Epic electronic health record system. Researchers found that nearly two-thirds (62%) had adopted an ambient AI documentation tool by 2025. The three most adopted tools used by more than 80% of hospitals were DAX Copilot, Abridge, and ThinkAndor.

By allowing clinicians to focus more on patient care and reducing documentation burden, these tools may improve patient-provider communication and mitigate physician burnout. "Early evidence, including by Adler-Milstein and colleagues in this issue of AJMC, suggests ambient AI can meaningfully reduce documentation time—but our findings raise an important equity question," says Ilana Graetz, lead author and professor of health policy and management. 

Their findings show that adoption was uneven across the included hospitals. It was more common among hospitals with stronger operating margins, larger size, metropolitan location, nonprofit ownership, and higher staffing-adjusted workload. Geographic variation was also a factor, with lower adoption rates in the Midwest compared with the South.

“In providing the first national estimates of ambient AI adoption in US hospitals and identifying organizational correlates, our findings reveal uneven adoption patterns that could have important equity implications. If these technologies prove effective at reducing burnout and improving care quality, unequal access could widen existing disparities between well-resourced and under-resourced hospitals,” says Graetz.

The authors note that policy efforts, cost-effectiveness analysis, and implementation support—similar to those used to accelerate electronic health record adoption—may be needed to ensure that new AI technologies benefit the entire health care system.