Race-determining AI

Doctors cannot determine a person’s race from medical images such as x-rays or CT scans. But a team including MIT researchers was able to train a deep learning model to identify patients as white, black, or Asian (according to their own description) simply by analyzing such images, and they still can’t figure out how a computer does it. It.

By examining variables, including differences in anatomy, bone density, and image resolution, the research team “couldn’t come close to identifying a good proxy for this task,” the researcher says. paper co-author Marzieh Gassemi, PhD ’17, associate professor at EECS and the Institute of Medical Engineering and Science (IMES).

The researchers say this is a concern because doctors are using algorithms to help make decisions, such as whether patients are candidates for chemotherapy or an intensive care unit. Now, these results increase the likelihood that the algorithms “take into account your race, ethnicity, gender, whether you’re incarcerated or not — even if all that information is hidden,” says co-author Leo Anthony Cheli, SM ’09, director. IMES Fellow and Associate Professor at Harvard Medical School.

Seli believes that clinicians and computer scientists should look to sociologists for information. “We need another panel of experts to weigh in and provide information and feedback on how we design, develop, deploy and evaluate these algorithms,” he says. “We also need to ask data scientists before we examine the data: are there inconsistencies? Which patient groups are marginalized? What are the reasons for these differences?

Algorithms often have access to information that humans don’t, which means experts must work to understand unintended consequences. Otherwise, there is no way to prevent algorithms from perpetuating existing biases in health care.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button