Published on
A national survey of 2,174 U.S. acute care hospitals found growing uptake of generative artificial intelligence (AI) tools integrated with electronic health records (EHRs) in 2024, as published in JAMA Network Open. A total of 762 hospitals (31.5%) were considered “early adopters,” meaning they currently use integrated, generative AI. Another 24.7% say they plan to have implementation within a year (“fast followers”), suggesting that more than half of the hospitals surveyed would have adoption by the end of 2025. Implementation was strongly associated with prior experience using predictive AI. Hospitals with predictive AI integrated into their EHR were 26 percentage points more likely to be early adopters or near-term adopters than those without predictive AI. Use was also higher among hospitals using Epic systems, large teaching hospitals, and those with stronger financial margins.
The digital divide: In unadjusted analyses, independent hospitals and critical access hospitals were less likely to be either early adopters or fast followers. Additionally, 20% of hospitals had a high proportion of Medicaid discharges, and these hospitals were 4.9 percentage points less likely to be early adopters or fast followers than hospitals with few Medicaid discharges. With usage becoming inevitable in all care settings, simply banning these tools can backfire and create liability; learn how to implement a safe, effective governance framework from the JUCM archive: The Case for a Generative AI Acceptable Use Policy in Urgent Care