Born and raised in Ethiopia, Timnit Gebru fled an outbreak of political violence in her homeland and sought asylum in the United States at the age of 16. As she recalled in an interview, “When I was growing up in Ethiopia, I sort of thought about racism in the U.S. I knew it existed. But I was not prepared at all for the way it was gonna happen. I just was completely shocked because a lot … of it is very direct.” She realized very quickly that she to advocate for herself and others to overcome bias both in and out of the classroom. Gebru would go on to earn a PhD from the Stanford University’s Artificial Intelligence Laboratory. While still a PhD student, she cofounded Black in AI, an organization fostering collaboration and initiatives to increase the representation of Black people in the field.
Gebru would go on to be named the coleader of Google’s Ethical Artificial Intelligence research team, where she worked to highlight biases and ethical risks within the technology and within the company creating it. She was eventually fired by Google for coauthoring a paper examining racial discrimination and bias present in large language models, a type of AI software. As she put it, “I’m not worried about machines taking over the world. I’m worried about groupthink, insularity, and arrogance in the AI community. If many people are actively excluded from its creation this technology will benefit a few while harming a great many.” Gebru’s departure instigated a labor movement and the creation of the first union for tech workers at Google. She now leads the Distributed AI Research Institute (DAIR), which documents the effect of artificial intelligence on marginalized groups. As she remarked in an interview with the Ford Foundation, “There is no such thing as a neutral or unbiased data set. So, what we need to do is make those biases and make those values that are encoded in it, clear.… People think that just because they’re putting data from the Internet, that everybody's point of view is represented. And that’s just not true, right? So what we need to do is make those values explicit.”
Twitter: @timnitGebru
