Smart CCTV Networks Are Driving an AI-Powered Apartheid in South Africa

Michael Kwet is a Visiting Fellow of the Information Society Project at Yale Law School. He is the author of Digital Colonialism: US Empire and the New Imperialism in the Global South, and hosts the Tech Empire podcast. "Beggars" and "vagrants" are not welcome in Parkhurst, South Africa, a mostly white, middle-class suburb of about 5,000 on the outskirts of Johannesburg's inner city. Criminals are on the prowl, residents warn, and they threaten their neighborhood security. To combat crime, the locals came up with a solution: place CCTV surveillance cameras everywhere. However, these are not the camera networks of times past. Thanks to advancements in machine learning and AI, CCTV systems are now equipped with sophisticated video analytics that can track a wide range of behaviors, objects, and patterns, in addition to individual faces. Armed with powerful new tech, communities of color can be watched, flagged, policed, and intimidated into submission. I've spent the past several years studying the video surveillance industry in South Africa. During that time, a private corporation called Vumacam has been quietly assembling a "smart" CCTV surveillance network in the suburbs of Johannesburg. Earlier this year, the company announced it would blanket Joburg with 15,000 cameras.