This August, tens of millions of Africans will turn their attention to Kenya’s general election. Kenya’s recent history features hotly contested, sometimes violent elections in which candidates and their allies have used tribal politics to turn people against one another. Yet as this election approaches, one of the biggest dangers comes much further from home: US and Chinese tech platforms.
Before unpacking the dangerous role that tech platforms are playing in Kenya, it’s important to understand the high stakes. For many Kenyans, this is the mother of all elections. The country’s president, Uhuru Kenyatta of the Jubilee party, has overseen an economy battered by inflation and debt, bruised by corruption and struggling to get on its feet due to the harm inflicted by Covid. Fighting to be the next president are Kenyatta’s deputy, William Ruto, and the leader of the opposition, Raila Odinga, of the Orange Democratic Movement (ODM). The last time Kenyatta and Ruto were on opposing sides of an election, in 2007 and 2008, the country was plagued by violence, and they ended up on trial in the international criminal court (ICC). (Kenyatta’s charges were dropped in December 2014, and the court terminated Ruto’s case due to weak evidence.)
Essential discussions about the election are unfolding on platforms such as Twitter, Facebook, and TikTok. It’s on these platforms that crucial civic information – but also disinformation and hate speech – will be amplified. Meanwhile, Kenya receives just a fraction of the resources – if that – that platforms give to address similar issues in western elections. They have acquired a massive civic responsibility in our countries – one that they are having trouble accepting.
These platforms already have an unpleasant history of abetting election disinformation across Kenya. In 2013 and 2017, Kenyatta’s campaigns used Cambridge Analytica to create election branding and messaging that critics called “divisive propaganda” that inflamed ethnic tensions. Similarly, Google ran poisonous attack ads under the banner “The Real Raila” on its search and YouTube products from Harris Media – a Texas-based rightwing media company that was also hired by Trump during his 2016 campaigns.
Now it’s happening all over again. In the past few months, work I published with Mozilla reported how Twitter’s complacency has allowed for the development of a thriving disinformation industry in Kenya, comprised of influencers-for-hire who sell their services on the platform to politicians and political groups. This industry has repeatedly been used by various problematic actors to consolidate power and neutralise public outcry, from lawmakers (or their proxies) to political groups abroad.
For example, in 2021 Kenyan judges and activists underwent wave after wave of attacks on Twitter as Kenyatta and Odinga sought to get their elite pact, titled Building Bridges Initiative (BBI), past the courts. Further investigations found that the same shady industry attempted to neutralise public outcry after Kenyatta was implicated in Africa Uncensored’s Pandora Papers revelations (along with other news organisations worldwide) in late 2021. (Kenyatta said the papers “would go a long way in enhancing financial transparency required in Kenya”, but never responded to his presence in the documents.)
It doesn’t stop there. Rightwing political organisations in places such as Spain have been using platforms to meddle with Kenyans’ civic debate online. The Spanish organisation CitizenGO ran campaigns over the past few years that Twitter amplified through its trending algorithm to millions of Kenyans in an attempt to oust politicians who support progressive legislation.
Meanwhile, allegations about Facebook’s moderation efforts in Africa dim any optimism that big tech companies might learn from past failures. Time recently reported on the alleged poor pay and working conditions of Facebook’s content moderators in Kenya. Moderators claimed that speed was prioritised over all else, including health. There are many criticisms that Facebook content moderators around the world are treated poorly, but the low pay and poor conditions in Africa are especially stark. African moderators accused Meta and Sama (Meta’s moderation subcontractor) of discrimination and rights abuses after the working conditions left many of them with PTSD.
Since the revelations came to light Meta and Sama have been served with legal letters for breaking multiple sections of Kenya’s labour laws. Such a situation means that hate speech and incitement will continue to run rampant on the platform.
Disinformation has been allowed to fester in east Africa, especially when compared to regions like North America and western Europe. This is evidence of two larger problems. The first is US tech platforms’ context bias in Africa. US platforms and the people who run them – most of whom are based in California – simply don’t know the histories and norms of African democracy.
Secondly, platform moderation and policies guiding AI regulation are enabling colonial interference, culturally and politically in digital spaces. As the Kenyan researcher and political analyst Nanjala Nyabola asks in reference to Cambridge Analytica’s hand in the Kenyan election: “What does accountability for political misinformation look like when a British company uses an American platform to influence political discourse in a Kenyan election?”
There are few incentives for platforms to address these problems. Big tech’s choices are consistently driven by public perception, business risk, the threat of regulation and the spectacle of PR fires. That’s why platforms are quick to scrub QAnon content in the US or meet GDPR regulations in the EU, but don’t show the same attention when dealing with health disinformation in Kenya, or adhering to the country’s incitement laws.
The policies and values of these platforms have normalised a kind of deviance – one that enables a dismissal of regions and populations that fall in its “rest of the world” category.
As the election draws nearer, many platforms are still unwilling to publicly commit to a roadmap that outlines how they are going to fight misinformation and disinformation in Kenya and Africa more broadly. We need platforms to inform Kenyan users about how they will use algorithms to spot hate speech and election-related disinformation; how they will foster relationships with civil society to factcheck content in English, Kiswahili and Sheng; and finally, how they’ll help Kenyans get accurate information about where to vote.
Election disinformation is not a problem that platforms can fix on their own. However, they have an outsized responsibility to protect Kenyan civil society and our democratic discourse.
Odanga Madung is a Mozilla fellow, journalist and data scientist based in Nairobi, Kenya
