Chest X-rays analyzed by AI could determine the race of the patient, with no additional data, and researchers don’t even know how the technology figured it out. (Source: AdobeStock)

New Capability Raises Spector of Racism in Medical Setting Based on Chest X-rays

The existence of discrimination in healthcare treatment for black patients is a fact, as confirmed by multiple studies. But the idea that technology could help eliminate that differentiation in treatment has been proven to be untrue. In fact, researchers recently discovered that AI could tell one race from another simply by looking at X-rays.

A research project shared by a large team of researchers from different institutions found that AI could tell a patient’s race by reading chest X-rays—or even just parts of them. The research was published in the online medical journal The Lancet Digital Health.

“We aimed to conduct a comprehensive evaluation of the ability of AI to recognize a patient’s racial identity from medical images,” write the researchers in their published paper. We show that standard AI deep learning models can be trained to predict race from medical images with high performance across multiple imaging modalities, which was sustained under external validation conditions.”

According to a story on, “The findings raise some troubling questions about the role of AI in medical diagnosis, assessment, and treatment: could racial bias be unintentionally applied by computer software when studying images like these?” The international team of health researchers from the U.S., Canada, and Taiwan tested their system on X-ray images the computer software hadn’t reviewed previously.

“The AI could predict the reported racial identity of the patient on these images with surprising accuracy, even when the scans were taken from people of the same age and the same sex. The system hit levels of 90 percent with some groups of images.”

The implications are giving researchers and medical ethicists some pause and concern about how to deal with AI that could be used to discriminate.