The doctor cannot determine whether a person is black, Asian or white just by looking at their x-ray. But the computer can, according to an amazing new work by an international team of scientists, including researchers from the Massachusetts Institute of Technology and Harvard Medical School. From the report: Research has shown that an artificial intelligence program prepared to read X-rays and CT scans can predict a person’s race with 90 percent accuracy. But scientists who conducted the study say they do not know how the computer calculates. “When my graduate students showed me some of the results that were in this article, I thought it was probably a mistake,” said Marzi Gasemi, an associate professor of electrical engineering and computer science at the Massachusetts Institute of Technology and co-author of a paper published Wednesday in the medical journal. The Lancet Digital Health. “I honestly thought my students were crazy when I was told.”
At a time when AI software is increasingly being used to assist physicians in making diagnostic decisions, research raises the alarming prospect that AI-based diagnostic systems may inadvertently generate racial outcomes. For example, artificial intelligence (with access to X-rays) can automatically recommend a specific course of treatment for all black patients, regardless of whether it is best suited to a particular person. Meanwhile, the patient’s doctor would not have known that AI based his diagnosis on racial data.