Doctors cannot tell if a person is black, Asian or white just by looking at X-ray imaging results. Computers can , an international team of scientists including researchers at MIT and Harvard Medical School says in a surprising new paper. An artificial intelligence program trained to read X-ray imaging results and CT scans can predict a person’s ethnicity with 90 percent accuracy, the study found. But the scientists who conducted the study said they didn’t know how the computer did it. “When my graduate students present some of the results in the paper, I actually Thinking it must be a mistake.” “I really thought the students were crazy when they told me.” The study comes at a time when AI software is increasingly being used to help doctors make diagnostic decisions. A troubling prospect is that AI-based diagnostic systems could inadvertently produce racially biased results. An AI with access to X-ray imaging results, for example, might automatically recommend a specific treatment for all black patients, regardless of whether it’s best for a particular patient. At the same time, the patient’s human doctor did not know that the AI’s diagnosis was based on racial data.
This article is reprinted from: https://www.solidot.org/story?sid=71545
This site is for inclusion only, and the copyright belongs to the original author.