This South African Researcher Is Using AI And Satellite Images To Tackle Spatial Apartheid
AI and satellite imagery are being used to fight spatial apartheid in South Africa, exposing stark inequalities and driving systemic change.
According to MIT Technology Review, Raesetje Sefala, a 28-year-old researcher from Limpopo, South Africa, is at the forefront of using AI and satellite imagery to combat spatial apartheid.
Growing up in a township, Sefala witnessed firsthand the disparities between Black townships and affluent white neighborhoods.
Now, she’s using cutting-edge technology to address these injustices.
Unearthing Spatial Inequities
In collaboration with computer scientists Nyalleng Moorosi and Timnit Gebru at the Distributed AI Research Institute (DAIR), Sefala’s team has developed a comprehensive dataset using millions of satellite images and geospatial data from the South African government.
This dataset maps out townships, allowing researchers to analyze changes in population and neighborhood sizes.
The ultimate goal is to identify and address the persistent inequalities in resource allocation and public services between townships and wealthier suburbs.
Sefala’s work has revealed that over 70% of South African land remains vacant, highlighting the disproportionate land allocation to townships versus suburbs.
By sharing this data with researchers and public service institutions, DAIR aims to facilitate the identification of land that could be repurposed for public services and housing.
Impact on Policy and Advocacy
One significant application of this data is its use by the Human Sciences Research Council (HSRC) to advise the government on budget allocations for health services, including HIV treatment programs.
Accurate mapping of townships and their growth rates enables more realistic resource distribution, ensuring marginalized communities receive the support they need.
Moreover, this initiative supports organizations like Ndifuna Ukwazi, which advocates for urban land justice in Cape Town.
Feature Image Credit: Hannah Yoon
