The older Sefala became, the more she peppered her father with questions on the visible racial segregation of their neighborhood: “Why is it like this?”

Now, at 28, she helps do something about it. Alongside computer scientists Nyalleng Moorosi and Timnit Gebru on the nonprofit Distributed AI Research Institute (DAIR), which Gebru arrange in 2021, she is deploying computer vision tools and satellite images to investigate the impacts of racial segregation in housing, with the final word hope that their work will help to reverse it.

“We still see previously marginalized communities’ lives not improving,” says Sefala. Though she was never alive in the course of the apartheid regime, she has still been affected by its awful enduring legacy: “It’s just very unequal, very frustrating.”

In South Africa, the federal government census categorizes each wealthier suburbs and townships, a creation of apartheid and typically populated by Black people, as “formal residential neighborhoods.” That census is used to allocate public spending, and after they are lumped along with richer areas, townships are in effect hidden, disproportionately excluding the people living there from access to resources resembling health services, education centers, and green spaces. This issue is usually referred to as spatial apartheid. 

Sefala and her team have spent the last three years constructing a knowledge set that maps out townships so as to study how neighborhoods are changing when it comes to population and size. The hope is that it could help them see whether or not people’s lives in townships have improved because the legal dissolution of apartheid.

They did it by collecting thousands and thousands of satellite images of all nine provinces in South Africa, and geospatial data from the federal government that shows the placement of various neighborhoods and buildings across the country. Then they used all this data to coach machine-learning models and construct an AI system that may label specific areas as wealthy, non-wealthy, non-residential, or vacant land. 

In 2021, they found that over 70% of South African land is vacant, and so they saw how much less land is allocated to townships than to suburbs. It was a confirmation of the inequalities they’d expected to see, however the staggering quantity of vacant land still took them aback, says Sefala.

Now they’re sharing the information set with researchers and public service institutions, including nonprofits and civic organizations working to discover land that may very well be used for public services and housing. DAIR plans to make the information let loose and accessible on its website from February 2.