Influence Operations And Disinformation In A Digital Age
“We always knew that this election was going to be fought on tech.” – the worlds of a prominent Kenyan online influencer who used his significant online following, ability to “game” social media algorithms and a variety of inauthentic amplification techniques, disinformation and even subterfuge, to ensure that his candidate got maximum exposure
online during Kenya’s 2022 Presidential election campaign.
In Kenya’s rapidly expanding digital environment, the weaponization of information online was a tactic observed by all sides in last year’s election – a contest which saw William Ruto pitted against former President Uhuru Kenyatta’s favoured candidate, the veteran politician Raila Odinga. Our study for the African think tank the Institute for Security Studies (ISS) “A question of influence? Case study of Kenya elections in a digital age” which used data analytics, OSINT and field research and examined some 8 million online documents, highlights the extent towards online influence has become a commodity – in many instances unrelated to ideology or conviction.
Furthermore, it shows that although in Kenya’s case, influence campaigns were largely home grown, the same tools and techniques could in future be used by other actors including hostile nation states, transnational criminal networks and terrorist organiations to shape public debate, political narratives and undermine democratic institutions – at scale.
With elections on the horizon in other digitally advanced African states – most notably South Africa in 2024 – Kenya’s experience of influence operations and disinformation, could offer some important lessons.
There are more than 11.8 million users of social media in Kenya, a figure which has grown three-fold since 2014 according to Statistica.. While increased access to digital technology have helped to advance freedom of speech and robust political debate online, it has also provided a platform for coordinated inauthentic activity designed to distort, intimidate and essentially “hack” the electorate.
Why does this matter? It matters because some users of social media during election time, (including some newsrooms,) may find themselves used as unwitting foot soldiers to amplify disinformation or conspiracy theories online. It matters because one of the key pillars of any democratic state – its media and legacy or traditional media in particular – find themselves unable to compete and verify, contextualise or correct, highly coordinated influence operations (often masquerading as “news” ) in an information environment where speed is king.
During Kenya’s Presidential contest last year, fact checking organisations did an admirable job of trying to call out disinformation and inauthentic activity. Much of this activity was aimed at confusing voters or voter suppression i.e., frightening them into staying away from the polling stations in crucial constituencies. But fact checkers’ can only do so much in a limited time and as we head into an era of ever more advanced AI, increased automation is likely to outpace the human ability to verify information in a timely manner. In turn that paves the way to for an automated disinformation arms race, in which Africa’s most vulnerable digitally connected democracies are most vulnerable.
Our ISS research throws light on a highly lucrative market for influence and hashtags for hire in Kenya, largely driven by commercial gain rather than ideological conviction. Product influencers lend their networks to political actors in return for significant financial rewards, only to return to promoting products after the election dust has settled.
The study explores the market for influence in granular detail, setting out the prices and ecosystem of online influence during election time in Kenya. It suggests that the rise of political “influence as a service” enables a highly skilled cadre of individuals, to tap into existing fears, social cleavages and conspiracy theories, and amplify them online. Moreover, these activities often occur without the users of social media being aware that they are co-opted as cogs in a complex wheel of influence and often distortion. Simply by liking, sharing, or re-tweeting a piece of content, they may be inadvertently widening the dragnet of disinformation.
Kenya is no stranger to influence operations. Remember the online work of Cambridge Anaytica which used Kenya for a carefully crafted campaign which tapped into ethnic prejudices and fears of Kenyan voters? Kenya was a perfect testing ground for such social media manipulation, ahead of the 2016 US Presidential elections in which the company also boasted its involvement.
Now with more of the Kenyan electorate online, the possibility of shaping domestic narratives at speed and at scale is vast. Kenya positions itself as a supporter of freedom of speech and has a proud tradition of a highly politically engaged population and has a highly educated cadre of digital marketeers. Yet, the prospect of Kenya outsourcing its own influence expertise to potential malign actors across the region and even globally, who want to do more than simply push products or merchandise, should be monitored by researchers.
Furthermore, outside of an election setting, for instance during times of national crises, the tools and tactics observed in Kenya in future may be exported and mirrored elsewhere whether it is to justify xenophobic narratives, harden attitudes against minority communities, or give legitimacy to coups. Wagner is doing this already. We can expect more players in this lucrative influence market presenting a real challenge to the principles of democracy.
On 8th August 1100 – 1230 South Africa time (UTC +2) ISS will host an in person and online discussion as a panel presents the Kenya research. The event will be moderated by the prominent South African journalist Ferial Haffajee and participants and panelists will investigate the lessons learnt for other African states preparing to hold elections, looking at South Africa in particular as it prepares to go to the polls in 2024.