Hi Quartz Africa readers,
I spent the better part of this week reflecting on why artificial intelligence (AI) treats Africans and people of color unfairly. First, I searched for “ear” on Google Images and almost all results were of white skin tones. I repeated the search with “leg” and only white legs appeared.
Tech has historically been a very white, male-dominated space and data representing white and male populations—rather than the wider user base—have become the default in building machine learning models. People of color continue to experience internet bias because global tech giants don’t have enough senior staff fighting for inclusion in the company’s AI algorithmic training.
Even those who attain leadership roles in the industry can struggle to achieve results. For example, Ethiopian-born US AI scientist Timnit Gebru believes she was pushed out of her role as…