High-profile Twitter users in Kenya are being hired by a mysterious agency to participate in highly organized harassment and disinformation campaigns targeting activists and members of the East African country’s judiciary, a new report from the Mozilla Foundation has found.
For $10 to $15 (about £7.25 to £10.90) a day, influencers allegedly peddle unfounded claims, malicious content and tweet propaganda material under designated hashtags – turning the cogs of “well-oiled” disinformation campaigns designed with the intent of swaying public opinion during “high pressure” political events, maligning activists, and undermining trust in Kenya’s judiciary.
With help from an army of bots, rented verified accounts, and sock puppet accounts, the report claims harmonised disinformation campaigns play on Twitter’s algorithm to the platform’s “Trending” section. But they don’t know where the money is coming from.
It was when Kenya underwent a hotly contested constitutional review process popularly known as the Building Bridges Initiative (BBI) – a scheme to make fundamental changes to the structure of the East African nation’s constitution – that Brian Obilo and Odanga Madung, Mozilla fellows and authors of the report, noticed that something was wrong.
“Only a given number of hashtags trend in Kenya in one day,” Madung told VICE World News. “When we looked on Twitter’s ‘Trending’ section, we noticed a certain type of hashtag would trend. When we looked at the semantics of the hashtag campaign, we noticed that it would either be attacking members of the judiciary, promoting the constitutional amendment bill (BBI), or attacking prominent figures and dissenting voices that opposed it. We saw hashtags shoot up for one day and then disappear. Just one huge spike…and then flatline. Organic Twitter activity does not work like that: and these are the tell-tale signs of inauthentic coordinated behaviours which we saw were being used for the purpose of attacking members of civil society.”
Tracking hashtag campaigns related to the BBI between May and June of 2021 – the height of the landmark constitutional reform ruling – Madung and Obilo uncovered evidence of at least 11 distinct disinformation campaign operations consisting of over 23,000 tweets, sent from 3,742 unique Twitter accounts. At least 31 artificial political hashtags, including those related to the constitutional amendment bill, were identified by the Mozilla Foundation: meaning that Kenyans on Twitter were being exposed to “at least one disinformation campaign every two days”.
“The agility of the tactics is in how well organised these guys can be,” says Madung. “We found that anonymous organisers used WhatsApp groups to send influencers detailed briefs which gave instructions on the material they would be posting, the predetermined hashtags that would accompany their tweets, and details of who they would be targeting for the day. They would then draft a plan to synchronise their activity before executing on the Twitter campaign and receiving their payment directly via M-Pesa [mobile money service]. To avoid detection, influencers would be sent their briefs individually – sometimes even renting verified ‘blue check’ accounts with large followings to execute their disinformation campaigns on.”
Mozilla researchers found that numerous verified accounts participated in identified disinformation campaigns used to attack Kenya’s Judiciary – #AnarchistJudges and #Justice4Sale.
“There was also another hashtag that arose called #WakoraNetwork which translates to ‘Network of Crooks’ in Swahili,” Obilo told VICE World News over the phone. “When the judiciary nullified current President Uhuru Kenyatta’s victory in 2017, he referred to the judiciary as ‘Wakora’. After judges ruled that the BBI constitutional amendment ruling was “illegal and unconstitutional”, the same hashtag resurfaced again. As you can see this is not something that started recently, it is something that also happened in the past.”
The research shines fresh light on the silently booming shadow industry of disinformation-for-hire in Kenya: a microcosm of a burgeoning global disinformation industry run by private firms who are estimated to have pulled in almost $60 million since 2009. The Oxford Internet Institute warns that Kenya – which has been ranked as an African nation with “medium cyber troop activity” alongside Nigeria, Rwanda and Ethiopia – is witnessing a “steadily growing” increase in political actors using social media to “disrupt elections, democracy, and human rights.”
Attribution back to particular actors, however, “remains difficult.”
The BBI is not the first Kenyan political event to have left a slew of organised social media disinformation campaigns in its wake. In 2018, it was revealed that UK consultancy firm Cambridge Analytica was alleged to have acquired Facebook user data of Kenyans, before targeting them with attack ads and misinformation without their consent to sway the outcome of Kenya’s highly contentious 2017 elections. In an investigation by the UK’s Channel 4 News, Mark Turnbull, managing director of the British firm, was filmed boasting about the control his firm exerted on Kenya’s 2013 and 2017 elections: admitting to having had a hand in “just about every element” of current Kenyatta’s presidential campaign.
Obilo tells VICE World News he was shocked at the number of sock puppet accounts found amplifying disinformation undetected on Twitter during the BBI review: “While analysing the accounts participating in the hashtag campaigns we spotted so many sock puppet accounts,” Obilo says. “One thing that came out very clearly that showed how shocking this was was that some of these accounts were not only created during the same time period, but they were using stock photos for their profile picture. And we are not talking of just one account – we are talking of over 15 accounts that had the same profile picture.”
The Mozilla Foundation report found that sock puppet accounts also targeted prominent activists, such as those from the Linda Katiba (Protect the Constitution) movement, who vocally opposed the constitutional reforms proposed by the BBI. The attacks were acute, and found to be of a nature that “border[ed] on incitement”.
“The disinformation attacks against me focussed on painting me as someone with ulterior motives who isn’t interested in the welfare of Kenyans,” says Jerotich Seii, a Kenyan activist and member of Linda Katiba who was targeted by the organised harassment and disinformation campaigns. “I had to spend a good chunk of my time defending my position as someone who is actually a patriot who does what they do out of love for their country.”
Twitter has since removed over 100 accounts operating in Kenya which it found had violated its platform manipulation and spam policy. “While we weren’t able to independently verify the tweet-for-pay activity… we could confirm the presence of at least one network of coordinated accounts — which appeared to link back to an earlier set of enforcements against similar activity, carried out by our team in 2020,” a Twitter spokesperson said in a statement.
On the 20th of August, Kenya’s Court of Appeals upheld its judicial ruling – unanimously declaring the constitutional amendment bill “unconstitutional, null and void”. But with presidential elections around the corner, experts warn that if left unchecked, Twitter disinformation campaigns pose a serious threat to Kenya’s democracy.
“The biggest threat is that it pollutes the information ecosystem to an extent that it becomes so difficult for the citizens to separate the facts from the falsehoods,” Alphonce Shiundu, Kenya Editor of Africa Check, tells VICE World News. “The very makeup of Kenya’s digital community is that very few people use the internet – the 2019 census put this number at about 10 million – which works to about 1 in 4 Kenyans. When a disinformation campaign is carried out, there’s usually cross-pollination of the disinformation, say a screenshot from Twitter being shared on Facebook, WhatsApp and Instagram and such, and then it jumps off those platforms into the word-of-mouth grapevines, and before you know it, it is being repeated as fact in marketplaces.”
“Twitter’s current whack-a-mole approach [isn’t] working,” say Madung and Obilo. “Sustainable solutions that take into account the mechanics and incentives that drive these campaigns must be sought.”