Harvard University Celebrates Chepchumba Choge for Promoting Diversity by Reducing Algorithmic Bias

Kenyan student, Chepchumba Choge, joined Harvard University in the United States in 2018 and was ready to advance her career in computer science at the immersive learning institution. 

During her stint at the University, the Kenyan native encountered challenges that prompted her to pursue social justice on top of her computer science course. During her freshman year, a phone facial recognition technology mistook Choge for someone else in her class due to problems with the technology in identifying people of African descent. The problem stretched to affect more colleagues in the institution, a concern that elicited her urge to pursue joint degrees in computer science as well as African and African American Studies. 

“I’d only moved to the U.S. a couple of months before and was still coming to terms with my own racial identity as a Black woman. That case study made me think about how the biases that exist in the physical world are inadvertently getting passed onto our technology. I realized that algorithms are not neutral. They have the potential to propagate harm.” Choge said. 

This inefficiency in phone recognition convinced Choge to embrace an academic path that bridges technology and social justice. She has graduated with joint degrees in computer science and African & African American Studies, a rare combination at Harvard. During her senior year, Choge participated in the SEAS Office of Diversity, Inclusion, and Belonging (DIB) where she expressed her view of eliminating racial bias in technology by diversifying the workforce behind the technology. 

Havard celebrated her on Tuesday, May 16 for crafting a niche geared towards fostering a community of belonging for people who felt left out. Choge was also a member of several campus affinity and diversity groups including the African Students Association, Women in Computer Science, Harvard Society of Black Scientists and Engineers, and Society of Women Engineers.