In recent weeks and months, artificial intelligence has dominated worldwide news and has been deployed in basic daily life activities. AI refers to machines’ ability to perform functions such as learning, planning, thinking, and creativity.
People of color have long criticized AI, claiming unfairness and racial injustice. Deborah Raji, a Mozilla fellow and CS PhD student at the University of California, Berkeley who is interested in algorithmic auditing and evaluation, was drawn to this.
After her third year of college, Raji interned at the machine-learning firm Clarifai, where she worked on a computer vision model that would let clients flag problematic photos as “not safe for work.” She discovered, however, that it flagged photographs of people of color far more frequently than those of white individuals. She explained the imbalance as a result of the data training. According to Innovators Under 30, the model was learning to distinguish NSFW imagery from porn and safe imagery from stock photos.
Po.rn, it appears, is far more diverse, and this diversity was prompting the model to instinctively correlate dark complexion with deplorable content. When she informed Clarifai of the problem, the corporation refused to cooperate.
“It was very difficult at that time to really get people to do anything about it,” she recalled. “The sentiment was ‘It’s so hard to get any data. How can we think about diversity in data?’”
Raji refused to back down. She continued her research, looking into popular data sets for computer vision training. Her investigation revealed disturbing demographic discrepancies, as several data sets of faces lacked dark-skinned individuals. As a result, face recognition systems were unable to distinguish between such faces accurately. At the time, police departments and law enforcement organizations relied extensively on these technologies.
“That was the first thing that really shocked me about the industry. There are a lot of machine-learning models currently being deployed and affecting millions and millions of people,” she said, “and there was no sense of accountability.”
This led Raji to shift her focus away from the startup world and toward AI research, focusing on “how AI companies could ensure that their models do not cause undue harm—especially among populations that are likely to be overlooked during the development process,” she told TIME.
“It became clear to me that this is really not something that people in the field are even aware is a problem to the extent that it is,” she said to the outlet.
She is now concentrating her efforts on developing ways for auditing AI systems both within and outside of the firms that are developing them. She has also partnered with Google’s Ethical AI and the Algorithmic Justice League on the Gender Shades audit project. This experiment “evaluated the accuracy of AI-powered gender-classification tools created by IBM, Microsoft, and Face++,” Raji explained to TIME.
Her outstanding contributions to AI earned her a spot on Time Magazine’s inaugural list of the 100 most influential individuals in Artificial Intelligence (AI). She was assigned to the ‘thinkers’ group.
Raji was born in Nigeria’s Port Harcourt but relocated to Mississauga, Ontario, when she was four years old. Her family, she claims, left Nigeria to escape the country’s insecurity and provide a better life for her and her siblings.
Her family subsequently relocated to Ottawa, where she enrolled in college. Her family wanted her to become a doctor, thus she was interested in pre-med studies at the time. She was accepted to McGill University as a neuroscience major, but during a visit to the University of Toronto, she met a professor who convinced her to pursue a career in engineering.
She enrolled in her first coding lesson and soon found herself immersed in the world of hackathons. She soon discovered she could transform her thoughts into software that would aid in problem solving or system change.
