What you might see on Google today.
On Wednesday afternoon, Xolisa Dyeshana queried Google’s Algorithm with a simple request on Twitter: “Google ‘Squatter camps in South Africa’, go to images and see what happens.”
If you follow Dyeshana’s instructions, like many Twitter followers did, you will see a Google search that predominantly returns articles and images of poor white people.
The reaction on Twitter had many questioning how the search results are selected with some suggesting that the algorithm was pushing the “white genocide” narrative and undermining the land debate.
Some have accused the Afrikaner interest group AfriForum of being behind the search results.
But, as one Twitter user pointed, the Google Image search results work on an algorithm.
In 2016, the Mail & Guardian published an article, How does the Google Image search algorithm work – and is it racist?, which looked at how the algorithm does not intend to be racist, or biased, or abilist, or transphobic, but was actually designed to reflect conversations that happen online.
The algorithm has over 200 factors that are taken into account for page ranking from the age of the page, to the expertise of its coding, to even the number of images that are embedded in the article.
A site’s credibility and whether or not the site has been linked to “expected” websites — sites that have a connection to the industry — also play a role in the algorithm.
Google also gives preference to blog networks that are believed to be curated by real people, and not considered to be fake blogs.
So how does Google Images get flooded with images of white squatters?
It’s quite simple. Of the 200 factors that Google uses to determine its algorithm, traffic plays a key role.
The search engine uses data from Google Chrome to determine that quality sites are those that get direct traffic. A study conducted by Semrush in June 2017 also found that the time spent on a site, or one of its pages, the bounce rate, and keywords in the body and title all played a role.
One of the conversations that may have led to the plethora of white squatters on Google Images could have been the cause of an SABC article published in March this year which looked at a growing community of white quatters outside White River in Mpumalanga.
According to Twitter user @No_H_in_Ismail, the article had a high number of hits, which would explain why it would feature prominently.
Google does have some control over its algorithm that creates the biased search results, but the search engine has almost no control over the content that is created by individuals through websites, blogs or forums, and how well connected those sites can be.