Biased Algorithms Create Negative Health Outcomes for People of Color
Computers may not have biases, but the algorithms that run on them are created by people who do, whether it's explicit or not. Researchers from the University of California, Berkeley determined the effect these algorithms can have on the health of people of color. Their results showed that an algorithm used by one health care company provided results that said just 18 percent of African Americans it served were in need of more care, compared to 82 percent of those who were white. In reality, researchers said the “true reflection” was that 46 to 54 percent of African Americans the company served were in need of more care.
Illinois Tech Professor of Philosophy Elisabeth Hildt will moderate a virtual discussion—the second of a four-part series—that focuses on how algorithms like this impact the level of care people of color receive. Joining the discussion will be John Fahrenbach, a data scientist at UChicago Medicine; Stephanie Eaneff, a data science health innovation fellow at UC Berkeley's Institute for Data Science; and Chicago-Kent College of Law Visiting Assistant Professor Alexandra Franco.
You can view the full program on YouTube.Register