Send message to Author

Short abstract:

The paper offers an empirically grounded, and theoretically informed, account of how AI-powered urban surveillance in Africa builds on long-term histories of control and segregation, rather than being simply influenced by the latest advances by AI superpowers.

Long abstract:

Smart cities incorporate some of the most critical tensions introduced by advancement in AI, in relation to their social and political consequences. They can inform and improve the provision of services – e.g., through better traffic and transport management – but they can also be used as instruments of surveillance and control. Their duality has been a source of increasing concerns, not only in authoritarian states, accused to exploit them as ways to expand their gaze in return for better services (Hoffman, 2022), but also in more open societies.

This study compares two AI-powered law enforcements projects deployed in Johannesburg and Cape Town, in South Africa. The first, the Safe City project developed by China’s Huawei in Johannesburg, links cameras to form wide area networks that use artificial intelligence (AI) to index, sort and interpret data pooled into centralised surveillance-based “nerve centres”. The second, the Shot Spotter project implemented by US company SoundThinking in Cape Town, uses smart sensors and AI to help law enforcement agencies collecting evidence on gun shooting incidents and help detecting patterns of gun violence. The structured comparison of the two projects allows moving beyond simplistic dichotomies between supposedly democratic and authoritarian uses of AI in urban environments. It illustrates instead the complex interactions between the socio-technical imageries originating in the countries from which specific technological solutions originate (China and the US) and the longer histories of attempts by South Africa’s city administrations to seize digital technologies to fight crime and improve service delivery.