Defending information integrity in a digital democracy

Election season in Kenya takes over ordinary life. Campaigns swallow weekends, moving from churches to rallies to entertainment spots and, more than ever, to social media.

As the tools have grown sharper, so has the contest over what voters believe. Politicians and their allies can now shape opinion in ways that are far harder to track or contain.

Kenya's young, connected population is what makes the digital space matter. In the 2022 General Election, voters aged 18 to 34 made up about 40 per cent of the 22.12 million registered voters, according to the Independent Electoral and Boundaries Commission.

The audience has kept growing. Kenya had 23.4 million internet users at the end of 2025, and 18.4 million social media identities were recorded that October, according to DataReportal. Nearly 79 per cent of internet users were on at least one platform.

That reach lets politicians explain policy and meet voters where they are. It also makes it easy to invent stories, inflate development records and spread unverified claims about rivals.

Democracy at risk

Algorithms speed the process along, carrying emotionally charged posts to audiences already primed to share them. Social media can serve democracy, but misinformation and manipulation are raising hard questions about what it is doing to it.

Those questions drove a recent Freedom Cafe forum convened by the Friedrich Naumann Foundation Kenya, which brought together journalists, politicians, civil society groups, technology experts and government officials.

The foundation's country director, Ralf Erbel, said in his opening remarks that misinformation, hate speech and AI-generated content are deepening polarisation worldwide, including in Germany, and wearing down trust in elections.

Africa Check Country Director Alphonse Shiundu showed how creators dress up false claims as banter and how politicians use the same channels to push flattering narratives. Mainstream outlets add to the confusion, he said, when they publish conflicting or inaccurate reports.

The problem runs deeper than what any individual posts. Dr Shikoh Gitau, a Kenyan computer scientist, said the technology behind the platforms shapes what people see and believe. She cited research putting the cost of biased narratives to Africa at $4.2 billion a year.

"The AI tools know the information they have fed you, and they make sure the user's brain connects the dots in a certain way," she said.

Nerima Wako, executive director of Siasa Place, averred that civic groups feel it directly: platforms have flagged their voter education material while hate speech kept circulating. "Algorithms do not favour us," Wako said. "It is becoming clear that the big tech companies are thriving on toxicity."

The creator's role

That leaves creators carrying part of the weight. As audiences grow, so does the pressure to post things that land immediately.

Digital creator Tatiana Gicheru said the scramble for attention pushes people toward provocation. "For something to go viral, you have to tap into people's emotions," she said.

An IEBC representative recalled a claim from 2022 that a candidate had won a huge number of votes in Australia, which was not a diaspora polling location. An impossible number travelled anyway.

Artificial intelligence raises the stakes. Convincing text, images, audio and video now cost a fraction of what they once did, leaving voters less able to tell real political messaging from fabricated material.

Africa's turn

The same tools could give African countries more say over how they are seen. The continent has one of the world's youngest populations, and AI is projected to add as much as $1.2 trillion to its economy by 2030.

Dr Gitau urged Africans to shape the data these systems learn from, noting that live content, such as the streamer iShowSpeed's recent tour, is already denting old foreign assumptions about the continent.

"If we do not shape the narrative of Africa, the only thing known on the web will shape that narrative," she said.