/ 19 February 2024

Tech is already being used to manipulate us

Fake News Keyboard

Politicians are used to manipulating would-be voters through overhyped promises or outright threats – vote for me or lose your social grant.

It’s an area of constant innovation, from rallies in mega stadiums to billboard adverts and pamphlets handed out in minibus taxis. The new frontier is technology.

In South Africa in 2016, British public relations firm Bell Pottinger used an army of Twitter bots to stir up racial tensions. Their “white monopoly capital” campaign amplified the sentiment that white people in South Africa were hoarding wealth, while depriving black people of jobs.

Fake accounts would tweet hate. This would then make its way into society at large. And this distracted people from the clients of that campaign, brothers Ajay, Tony and Atul Gupta, who were working with then- president Jacob Zuma to systematically loot the country’s state entities, such as its already struggling energy utility, Eskom. The slick operation was eventually revealed and the fallout contributed to the collapse of Bell Pottinger.

Using this playbook, honed in the election of Donald Trump and the campaign to pull Britain out of the European Union, companies have been selling their manipulation skills across the continent. Their success is despite more than half of people in 38 African countries having no internet access.

Platforms like Facebook have been slow, or loath, to respond.

Ahead of the 2017 Kenyan election, automated bots on Twitter accounted for more than a quarter of the most influential accounts discussing the election, according to research by consulting company Portland Communications. Countries like Lesotho, Equatorial Guinea and Senegal saw a similar influx. The research found that these accounts had more of an impact than the accounts of politicians and their campaigns.

These “served primarily to agitate, pushing negative narratives about major issues, candidates, and perceived electoral abnormalities.” After the elections, many had their election content deleted, said the report.

Facebook employees and researchers at the Stanford Internet Observatory revealed in 2019 that a firm tied to the Russian mercenary Wagner Group was running a network of at least 73 Facebook pages targeting Africans.

The pages blasted over 8,900 posts, praising controversial figures like Libyan warlord Khalifa Haftar or flattering incumbents in Cameroon, the Central African Republic, Côte d’Ivoire, the Democratic Republic of the Congo, Madagascar, Mozambique, and Sudan. Facebook removed the pages but not before they had garnered over 1.7-million likes.

Ahead of Uganda’s 2020 election, Meta deleted 32 pages, 220 user accounts, 59 groups, and 139 Instagram profiles promoting President Yoweri Museveni. These focused on smearing support for his opponent Kyagulanyi Ssentamu (aka Bobi Wine) as “hooliganism”.

Facebook found that a few individuals at a government office and PR firms had used fake or duplicate accounts to manage these disinformation assets, a practice that the platform labels as “inauthentic coordinated behaviour”. Facebook was banned in Uganda for deleting the network and remains inaccessible without the use of a virtual private network (VPN).

In the DRC, people connected to political figure Honoré Mvula created fictitious personas to run 63 accounts on Facebook, as well as 25 Instagram accounts. The US-based Digital Forensic Lab reported that the Mvula network started as “pour le buzz” accounts – pages and groups impersonating celebrities or their fans. Once they had followers, they were renamed and their starter content was deleted.

Heading a “movement of young Congolese intellectuals,” Mvula campaigned for President Félix Tshiskedi in the 2018 election that brought him to power. Meta deleted the accounts in 2020.

In Kenya, with its vocal Kenyans on Twitter community, researchers at the Mozilla Foundation uncovered a network of 3,700 Twitter accounts which pushed at least 11 paid disinformation campaigns. Over two months in 2021, these blasted out over 23,000 tweets attacking journalists, civil society and public workers like judges.

In Nigeria in 2022, a BBC investigation found that political actors were secretly paying social media influencers as much as $43,000 to spread disinformation against their opponents ahead of the February general elections.

Because it is straightforward to create fake accounts and garner a following for them, campaigns like this are widespread. And this is before generative AI, which is predicted to become a big player in elections this year – 2024 will have more elections than any year in history.

As The Continent reports in an interview with Maria Ressa, the technology dramatically increases the capacity of groups that want to manipulate things like elections, and the regulation to stop this is falling behind.

This article first appeared in The Continent, the pan-African weekly newspaper produced in partnership with the Mail & Guardian, which is designed to be read and shared on WhatsApp. Download your free copy here.