AI Colonialism: A Case Study from Johannesburg
A 2022 article series from the MIT Technology Review has argued how artificial intelligence (AI) is repeating patterns of colonisation and colonial history. Rather than this occurring through the ‘violent capture of land, extraction of resources and the exploitation of people’, AI is doing so through data extraction. Its need for profit and domination makes it just as extractive as traditional colonisation.
Herein, AI represents a new mechanism of a historical system that once enabled powerful actors to dominate more than half the world often at the expense of poorer and more vulnerable populations. Especially in the Global South, new and cheap ways have been found to exploit communities. The article uses a case study from Johannesburg in South Africa to exemplify what this looks like.
In Johannesburg, networks of cameras are being used for private surveillance purposes, and are monitored by both humans and algorithms. This would not have been possible a number of years ago, but with the expansion of fiber coverage, companies abroad have started seeing opportunities in the Southern African country. The result of this is the ‘rapid creation’ of a centralised, coordinated privatised mass surveillance operation. The company building the CCTV network, Vumacam has nearly 7000 cameras, with more than 5000 of which are concentrated in Johannesburg. Access to the network operates through a subscription based model with roughly $50 a month being charged in 2019.
However, these cameras are only beneficial to those that can afford them. In South Africa, the most unequal country in the world and one shaped by a legacy of apartheid, this surveillance infrastructure is mostly used by the wealthier, often white communities for private security. Unfortunately, this imbalance can be traced back to apartheid South Africa where the National Party used the police to protect its political interests in the late 1970s, curbing widespread opposition to the government.
Nevertheless, the need for private security has ballooned due to the country’s staggering rates of crime - between 1st October and December 2025, South Africa recorded 6,351 murder cases. The cameras are located in front of homes, residential areas, schools, and malls to name but a few. Vumacam had pushed its solution as a way to track criminals from the moment they commit a crime to wherever in the city that they try to escape. Users can also add wanted vehicles to their own private database on the platform that enables all users to work together to track cars across jurisdiction. Some argue that this may be seen as facial recognition.
As a result, private security dominates the duties that are usually associated with the police in Johannesburg and greater South Africa, even though they do not have the same legal powers. A highly militarised private security sector has emerged with there being uniformed men in tactical vehicles holding big guns, a popular site around the city. Furthermore, instead of serving the public, they are hired by private citizens.
Vumacam’s growth has also been met with privacy concerns. However, in a country where inequalities are rife, more people are worried about immediate problems related to basic needs, employment, health, education, etc. Many of the technologies used in the country are also being developed with the US market in mind - the reason they are being used in South Africa is to give the company a competitive edge through cheaper local wages and to experiment.
Overall, the case study of surveillance in South Africa serves as an in depth example on the way that technologies influenced by AI are dominating. This is especially in countries and regions that are classed as ‘developing’. It illustrates how such technologies can reproduce unequal power dynamics.