Our Work

The MN-AIR Hub is an interdisciplinary team working to build the evidence base on AI and informatics to support digital transformation. This work includes original research and resource curation.

Spotlight

The National Opinion Research Center (NORC) at the University of Chicago conducted a survey on US adults aged 18 and older regarding their expectations of healthcare AI. 19.4% expect that AI will improve affordability of healthcare. 19.55% believe that it will improve their relationships with their doctors. 30.28% believe it will improve their access to care.
Data from hospitals was analyzed to see if they used and evaluated AI. 65% reported using AI or predictive models, but fewer than half systematically evaluated models for bias. Almost two-thirds evaluated models for accuracy. As AI adoption in healthcare increases, there is a digital divide in both implementation capacity and evaluation practices.
A 2024 study examined the adoption of generative AI within different hospitals' EHR systems. 31.5% currently used AI and 24.7% planned to adopt AI within the next year. Independent, rural, and high-Medicaid hospitals were less likely to adopt than their counterparts. Less than half of the hospitals reported conducting all model evaluations for bias, accuracy, and post implementation monitoring, and almost one-third of the hospitals indicated no local AI evaluation.
A study analyzed interviews focused on governance of AI-driven predictive models from individuals across Academic Medical Centers. Three AI governance phenotypes were identified: well-defined, emerging, and interpersonal. Even well-resourced academic medical centers reported a desire for clearer guidelines, resources, and rules that promote patient safety, care quality, and effectiveness of AI.
A national survey of US adults was analyzed to assess whether patients trust their health care system to use AI responsibly. Results showed most respondents had low trust, emphasizing the need for clear communication & stronger organizational trustworthiness.
A nationally representative survey of US adults found that there is an overwhelming desire for notification of AI use in healthcare. 95.2% of respondents thought that notification was important.