Study Finds That Facial Recognition May Be Key To Understanding Racial Bias

Researchers at Brown University have created a new training tool that they hope will help address unconscious racial bias through teaching subjects to better differentiate between faces of other races.

The study tested the racial bias of 20 Caucasian subjects by showing them pictures of a diverse group of Americans. After each face was shown, a word appeared on the screen, which could be either real or nonsense (like "tree" or "malk"). The test subjects were asked to say whether the word was positive or negative. The researchers found that there was an unconscious tendency in many of the participants to consider a word negative if it followed the face of an African American. After the trial, researchers put the subjects through hours of facial recognition training, followed by another round of faces and words. They found that the training worked on a number of levels: the individuals who showed the greatest improvement in telling apart the faces of members of other racial groups greatly improved in their word-associations, labeling more words positive after each non-white face was shown. They also showed a reduction in their implicit racial bias as measured by the ALPS system. The researchers hope that the facial recognition training program will someday be used to help train police officers, social workers, or immigration officials- anyone, really, who is at a high risk for exhibiting racial bias on a daily basis. [EurekAlert]