Proving it’s the data that is biased, not the algorithm through the recent ‘squatter camps in South Africa’ case study.

In the recent past, some South African internet users may have questioned the reliability of online news, but not the role of search engines play in programming public discourse. This changed in 2018, when South African Twitter users accused Google of peddling misinformation. Xolisa Dyeshana tweeted to his followers that a Google Image search for the term “squatter camps in South Africa”, predominantly displays images of white people in squatter camps. Dyeshana argued that by neglecting to display pictures of black people in squatter camps, it was serving a white agenda. These statements ignited a Twitter debate featuring disinformation, conspiracy theories and talk about media coverage. The topic trended on Twitter, was shared on other social media platforms and was picked up by mainstream news sites. Many analysts blamed Google’s algorithm for displaying bias. In this paper the authors use this case study to compare the findings on six different search engines, to counter this argument. Search engines that are diverse in their scope and origin are used to prove that is it not the search engine or the algorithm that is biased, but the data that is biased.

van Vuuren and Celik