Carolina: Artificial intelligence (AI) has virtually limitless applications in healthcare, from automatically generating patient messages in MyChart to improving the accuracy of organ transplants and tumor removal. includes improving
Despite their potential benefits to physicians and patients alike, these tools have been met with skepticism due to patient privacy concerns, potential for bias, and device accuracy.
A multi-institutional team of researchers from the UNC School of Medicine, Duke University, Elibank, Oxford University, Columbia University, and the University of Miami responded to the rapidly evolving use and adoption of AI medical devices in healthcare. It has a mission to build public trust and evaluate how AI technologies can be approved for use in patient care after sound clinical validation.
Sammy Choufani L. Fassi, an MD candidate at the UNC School of Medicine and research scholar at the Duke Heart Center, and Gail E. Henderson, a professor in the UNC Department of Social Medicine, conducted a comprehensive analysis of AI clinical validation data.
They found more than 500 medical AI devices approved by the US Food and Drug Administration (FDA), showing that almost half of the AI tools have not been clinically validated.
