Kenya’s recently concluded election has been lauded as transparent, credible and peaceful. However, in the closely contested presidential race between former Prime Minister Raila Odinga and outgoing Deputy President William Ruto, the torrent of misinformation and disinformation ran rampant online throughout the electoral process.
For millions of Kenyans social media has become the primary source of news, opening up spaces for political candidates to engage voters in more direct ways compared to traditional forms of campaigns such as rallies, billboard advertising and legacy media. During the electioneering period, different political factions through their “online armies” shared inflammatory posts and harmful narratives meant to discredit their opponents. According to a report by Mozilla Foundation, some of the most dramatic types of disinformation was spread on Tik Tok. Political jokes based on ethnic stereotypes that trivialise issues and could entrench animosity, messages of hate as well as speeches targeting other ethnic communities were freely conveyed on the platform. Odanga Madung, Mozilla Tech and Society Fellow who led the research said “a very clear goal of the disinformation apparatus is to erode trust, whether it’s in candidates, the electoral process or any other part of what it means to engage democratically.”
According to a report by the East African Newspaper, influential leaders from both sides of the two frontrunner coalitions shared manipulated media meant to tarnish the reputation of the opposing camps while boosting their own popularity. In one of such videos shared, subtitles are manipulated to depict ethnic incitement by Deputy President William Ruto. According to AFP Fact Check, the original footage of Ruto speaking at a political rally without any subtitles was of him pleading with the audience to be calm throughout the campaigning season and affirming that the area was home to people from all communities in Kenya.
Another was a fake audio reshared by a parliamentarian, claiming to be a leaked phone call between a high ranking official of Odinga’s coalition party, and an official of the electoral body attempting to rig the elections. Twitter flagged some of the fake media posted by both camps but not before they garnered traction that attracted outrage and condemnation.
Screenshots showing the leaders from both political camps accused of sharing manipulated media with the intention of framing their opponents:
Unverified declaration of results
As the complex process of tallying and verifying the presidential votes progressed slowly, before the final declaration by the electoral body, both the Odinga and Ruto camps were using social media to claim victory based on unofficial tallies. To exacerbate the situation, some influential and verified social media accounts published and declared the results in favour of the candidate they were supporting. In a joint statement by Amnesty International and other civil society organizations both political camps were cautioned against intentionally seeking to misinform the electorate and the public on the electoral process and the election results. Echoing the statement from the CSOs, observers from the Commonwealth and the US who also raised additional concerns about how the delay in releasing the final presidential results was an avenue for the spread of disinformation and stirring anxiety among supporters.
While Twitter flagged tweets bearing the misleading information as false, the labelling was inconsistent since it wasn’t applied to the hundreds of accounts that re-tweeted the same material.
Screenshots showing influential accounts from both political camps calling the final results of the presidential election:
Twitter also provided an election prompt in collaboration with the electoral commission where Kenyans will be able to access credible information about the elections when searching for keywords on Twitter, associated with the Kenyan General Election. Additionally, Twitter curated Moments designed to address misleading and false information around the presidential elections. The Moments included how the presidential results were to be announced from the electoral body and other credible information from media outlets, and fact-checking organizations:
In an attempt to frame negative public debate about some women politicians, false and misleading sex-based narratives against them were spread online. A campaign on Twitter meant to delegitimize women politicians associated with DP Ruto trended on Twitter under the pre-determined hashtag #Room350. It bore humiliating, sexually charged images and other sexual innuendos that implicated them. Equally, Odinga’s running mate Martha Karua was subjected to attacks based on targeting her physical appearance. Despite the harmful narratives against them, a record number of women were elected to various positions of power during this poll as the country strives to achieve gender equality. Twenty six female members of parliament, up from 23 in 2017, 7 female governors, up from 3 in 2017, and 3 female senators were elected.
Screenshots showing posts attacking the character and appearance of some women politicians:
The use of paid-for influencers to spread malicious and harmful content that targeted political opponents, journalists and their respective media houses was also on the rise. Those narratives were extended to the members of the civil society and activists who were discredited and portrayed as villains for holding brief for a particular presidential candidate. Jerotich Seii, one of the members of the Linda Katiba movement that pushed back against an attempt to amend of the Constitution of Kenya 2010, said in an interview that she had to spend significant time trying to prove that her activism efforts were genuine and not a front for someone else because of the attacks against her. One clear advantage the paid-for influencers gave their “clients” was lack of responsibility for the narratives. Other disinformation tactics witnessed were political astroturfing, mass brigading and the use of fake and pseudonymous social media accounts.
Screenshots showing coordinated mass brigading messages and pages on Facebook targeting both Raila Odinga and William Ruto:
Implications for democracy
While disinformation campaigns have become a feature of elections globally, they should be of utmost concern because of their serious implications for democracy including, fermenting politically motivated violence, fuelling tribal antagonisms and unsettling people’s faith in democratic institutions. Voters must have access to impartial, fact-based sources of information so they can form their opinions in the first place and then make informed political choices.
What social media platforms can do
Social media platforms are already tackling this challenge to some extent. In the lead up to the Kenyan general elections Twitter, Meta and Tik Tok all set guidelines to combat election related misinformation and disinformation but still struggled to control its spread. To test Facebook’s ability to detect on its platform ahead of the elections, Global Witness, an advocacy organisation and Foxglove submitted ten real-life examples of hate speech ads in English and Swahili to Facebook and both ad sets were accepted.
It is noteworthy that during the 2017 general elections in Kenya, Facebook was accused of failing to protect users’ personal data in the Cambridge Analytica breach. In the Mozilla Foundation report cited earlier in this article, Tik Tok too was named for failing to control the spread of “fast and far spreading political disinformation” on their platform. All these point to a broken system which requires effective responses at different levels. Experts have called on social media companies to demonstrate stronger corporate commitment to combat the spread of mis/disinformation on their platforms.
For years WITNESS has engaged with tech companies on content moderation, while a newer strand of our work is honed in on responses to digital mis- and disinformation. We have urged for greater stress to be placed on the human rights dimension of this challenge.
Here, we outline critical questions relevant to our work globally, and recommend actions moving forward:
- Ensure platform accountability is grounded in global realities and human rights. In the context of the Kenyan elections there were moderation gaps on Tik tok and Facebook due to lack of understanding of local language and context;
- Ground content moderation in human rights ensuring that counter-disinformation efforts do not amount to restricting the right to freedom of expression;
- Hold to account leaders who incite violence on social media and get away with it;
- Create workable ‘evidence lockers’ for preserving critical online content; and
- Defend witnessing.
- Media literacy campaign page
- Understanding the Dangers of Gendered Disinformation in African Elections
- Fact from Fiction: Curbing Mis/Disinformation in African Elections
- Misinformation and Disinformation in Elections Tipsheet
- How Manipulated Videos and Pictures Threatens Human Rights
- Tracing Trust: Why We Must Build Authenticity Infrastructure that Works for All
- WITNESS and the C2PA Harms and Misuse Assessment Process
Published 20th September 2022.