A recent article from TheSciVerse discussed the topic of racial bias in the context of medical imaging and artificial intelligence. Apparently an international team of researchers from the United States, Canada, and Taiwan tested AI on X-ray images and were able to predict the patient’s claimed race identification with up to 90% accuracy. The researchers think that it could be due to the system detecting melanin, the pigment in skin that gives it its color, in a novel way that is yet unknown.
"Our finding that AI can accurately predict self-reported race, even from corrupted, cropped, and noised medical images, often when clinical experts cannot, creates an enormous risk for all model deployments in medical imaging," the researchers wrote.
