Almost half of FDA-approved AI medical devices are not trained on real patient data, research reveals (Medical Xpress)

A study by researchers from institutions including UNC School of Medicine and Duke University found that around half of AI medical devices authorized by the FDA lack reported clinical validation data. Despite the growing use of AI in healthcare, particularly in diagnostics and treatment planning, the absence of thorough clinical validation raises concerns about the reliability of these tools. The study urges both the FDA and the industry to improve transparency and ensure that AI devices are rigorously evaluated using real patient data.