Skip to Main Content

Hospice nurse Deborah Carlson had been hesitant to embrace the artificial intelligence tool her health system introduced earlier this year. She knew next to nothing about its inner workings, its accuracy, or its reliability.

Then, it correctly predicted the death of a patient she thought seemed stable — and Carlson found herself questioning her judgement. How exactly did the tool work, and how could she know whether to trust it?


Health workers like Carlson are increasingly being asked to incorporate experimental AI systems into their workflows, as more hospitals roll out the technologies to predict patient needs and outcomes. But without a transparent explanation of how the tools work — or clear education on how to use them — health workers are forced to grapple with a range of unknowns. Nurses and doctors tasked with using the tools told STAT that the process has sowed confusion, frustration, and concern.

Unlock this article by subscribing to STAT+ and enjoy your first 30 days free!