In 2007, incumbent Mwai Kibaki won a divisive presidential election in Kenya. Street protests escalated to ethnic violence in parts of the country, and by April 2008 more than 1,500 people had been killed. A decade later, another election also featured widespread allegations of fraud, and more violence. The casualties were lower this time, but just over 100 were killed, almost all by the police in opposition strongholds.
Technology and politics are inextricably linked in Kenya, in part because technology was proposed as the solution to the structural issues that led to the violence following the 2007 election. The Independent Review Commission established in 2008 argued that technology would help bridge the trust gap between key political actors and protect the autonomy of the bureaucracy surrounding elections. In line with the commission’s recommendations, voter registration, voter identification, and vote tallying are all nominally computerized. These efforts culminated in the 2017 election, which was supposed to be Africa’s first fully digital election.
That they didn’t work is a lesson in what technology can and cannot fix in politics. Perhaps the primary lesson is that when communication platforms cater to specific audiences while shutting out others amid existing social divides, hate speech thrives. And this was as true of pre-internet technology as it is now.
Twenty-five years ago, Kenya had one state-owned radio and television broadcaster, but by 2017 there were over 60 licensed television stations and 178 radio stations. Kenya has two official languages—English and Kiswahili—and at least 44 ethnic groups, most of which speak a third language that’s often unintelligible to outsiders. As part of the democratization process in the 1990s, local-language radio stations were encouraged as way to bring more people into the national conversation while preserving regional cultures.
The unexpected result
But it turns out that local-language radio stations are particularly vulnerable to becoming channels for hate speech. They function as closed systems shielded from scrutiny by institutions that don’t have the capacity to manage them, or any interest in doing so. By the time the threat inherent in these stations was identified, it had already materialized, and regulators are still playing catch-up. Kenya has laws prohibiting hate speech in media and numerous bodies ostensibly working against it, but as late as July 2017, the National Integration and Cohesion Commission was warning that Kameme FM, a Kikuyu-language radio station owned by Kenyan president Uhuru Kenyatta, was responsible for the most incidents of hate speech related to the election. Despite this finding, the station received no public censure.
Technology has changed radically in Kenya over the past decade, as it has everywhere else. Almost nine out of 10 people have a mobile phone, and a quarter of the homes have an Internet connection—among the highest rates in the developing world. In a population of about 48 million, there are at least seven million Kenyan Facebook accounts and another 10 million on WhatsApp. Twitter lags behind at only a million accounts, but it is supplanting television and print as the premier space for political critique. mPesa, the mobile money transfer platform, was less than six months old at the time of the 2007 election. Today, transactions on mPesa equal almost a third of the country’s GDP, and Kenya has the highest number of mobile money transactions in the world.
A better way to manipulate
These technologies have transformed the way Kenyans produce and consume political information, and by extension the way they interact with each other. Yet the causes of hate speech on local-language radio stations haven’t gone away. Many of the practices horrifying analysts in the West—the exploitation of identity politics to win elections, or the influence of money on decision-making in electoral politics, for example—are familiar to Kenyans. (Cambridge Analytica, which has been accused of using Facebook data to manipulate voters in the US and UK, has been present and active in Kenya since at least 2012.)
Technology has given powerful people a more effective way of influencing the electorate by using language that dehumanizes the “other,” particularly on the basis of ethnicity, and then paying individuals to instigate violence that triggers reprisals. With the internet, information now travels faster, without the moderating impact of editing or verification. The same speed that makes social media the preferred avenue for alerting the public about emergencies makes it especially efficient at disseminating hateful opinions. Moreover, information on social media travels in relatively insular networks, which makes people less likely to encounter opinions that they disagree with or that challenge their thinking.
Division by design
According to experts like Zeynep Tufecki, some of the characteristics that encourage hate speech are embedded in the logic of these platforms, and indeed are what make them so attractive to advertisers; they are the features that make microadvertising and therefore profitability possible. Again, this did not begin with the internet; Kenya’s radio market, similarly, was stratified to sell advertising.
Stratified audiences may be good for advertisers, but they weaken the role of media because public discourse among different groups is no longer starting from the same point. Stratification also makes these spaces harder for regulators to monitor. In 2007, Kenyans rioted in response to allegations of violence, broadcast on local-language radio, that turned out to be unfounded. In 2017, ethnically charged memes alleged that a genocide would happen in Kenya should certain candidates prevail. The absence of fact-checking, verification, and oversight on these platforms, coupled with the speed of transmission, allows inflammatory rumors to spread.
Kenya reminds us that with every evolution of technology, new concerns arise about its role in politics. Technology reflects the values of the societies in which it’s deployed, and can’t fix problems that a society is unwilling to fix within itself. The internet has sped up political discourse and further insulated it from scrutiny, mimicking and amplifying the experience Kenyans had with local-language radio stations, and reminding us that some problems are bigger than the medium.
Nanjala Nyabola is the author of Digital Democracy, Analogue Politics: How the Internet Era Is Transforming Kenya.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.