An interim report released by the Centre for Analytics and Behavioural Change (CABC) has investigated, tracked and isolated accounts that are responsible for creating, supporting, and continuing xenophobic conversations on Twitter.
The report explained that xenophobic tweets function much like promotional accounts on Twitter, which is when accounts are used to promote products on a user’s timeline, in the hopes that the product will trend.
Users keen to continue spreading xenophobic content even create defunct accounts or pay other people to spread content, the report explained.
CABC decided to isolate sample accounts that were “inauthentic and had sings of planned manipulation” using a software programme called Brandwatch.
Criteria in isolating prominent accounts included tweet volume, political posts, connections, and a lack of personal tweets.
Brandwatch isolated the biggest accounts in terms of follower size and “leadership roles in the xenophobia conversation”.
They were identified as Lerato Pillay (@uLerato_pillay), Letshela Mofokeng (@SirLetshela), UmalambaneZN (@UmalambaneZN) and Patriot (@landback_).
Xenophobic tweets spiked significantly from March 2020, CABC determined and said that a number of accounts tracked down were dormant until November 2019, or March 2020.
“These two dates are significant because November 2019 was the inception of the @uLerato_pillay account and March 2020 marks the spike in the xenophobia conversation in South African Twitter,” the report explained.
CABC dug deeper and found that the Lerato Pillay ‘network’ has reached out to ATM president and MP Zungula Vuyo, former Johannesburg mayor Herman Mashaba, businessman Vusi Thembekwayo, and Mining Forum of South Africa president Blessings Ramoba. The report said these accounts also supported the hashtag #PutSouthAfricaFirst, which emerged on 27 April.
When it first emerged, it was used more than 16,000 times in one day. 80 accounts in the network around the Lerato Pillay account were responsible for 50% of the use of the hashtag.
“For another layer of evidence that these accounts may be connected, we looked at the content they shared.
“In most cases, there were similarities with the content shared, including pictures, videos, copy and captions. Most of these accounts are basically six posting the same content and retweeting each other.
“For example, the popular account would share a post, a week later one of the other smaller accounts would post the same tweet without even changing a word. In some weeks they will even use the same engagement methods, like creating a poll or a follow-for-follow trend.”
Brandwatch tracked the hashtag from 1 April to 31 May and found it was consistently identifying xenophobic content. In this time, the Lerato Pillay account has tweeted roughly 5,000 times, making it likely that more than one person runs the account.
The software also determined that a number of the individual accounts retweeting the hashtag identified as xenophobic on South African Twitter.
Further analysis revealed a map of networks with “clear evidence of strong connections between these accounts”. CABC concluded that this indicates they are “coordinating their activities.”
A list of suspicious accounts, once isolated, showed that the Lerato Pillay account was “deeply connected to the rest of this network”.
The Lerato Pillay account, CABC concluded, was imperative to the #PutSouthAfricaFirst hashtag, and is seen as the starting point for content that branches out into the rest of South Africa Twitter.
“This does not indicate organic growth in the conversation, but rather an organised network.”
On 1 April, there were roughly 12,000 people engaged in “organic conversation”, which grew to 50,000 in two months, and continues to grow every day. This means the 80 accounts are being managed to grow “a significant conversation by targeting foreigners living in South Africa”.
The report is a prominent step in identifying the origins of xenophobic content on social media.
Although CABC said a significant amount of content was organised “and probably funded”, the data did not account for the predicted large number of people that are part of the xenophobia conversation, but did not associate with the @uLerato_pillay account, or any other prominent instigators.