AI Disinformation Emerges as Major Threat to Elections in Africa and Europe

“Deepfakes, videos, audios — it’s so easy to produce disinformation at home for free with the help of artificial intelligence. What might happen in five or 10 years is really scary,” warns Hendrik Sittig, director of the Konrad Adenauer Foundation’s media program in sub-Saharan Africa.

The World Economic Forum’s 2024 Global Risk Report has identified AI-supported disinformation as the number one global threat, with campaigns often designed to undermine democratic principles and divide societies.

A comprehensive study conducted by Karen Allen from the Institute for Security Studies in South Africa and Christopher Nehring from Germany’s cyberintelligence institute, in cooperation with the Konrad Adenauer Foundation, has documented AI disinformation targeting electoral processes across Africa and Europe. Their research reveals that these campaigns primarily aim to undermine electoral authorities and processes, though they note that AI disinformation beyond election contexts remains largely uninvestigated in Africa.

The researchers discovered that Europe and Africa face remarkably similar challenges despite their different technological landscapes. “We could identify many culprits which use the same AI tools,” Nehring told DW, noting that extreme right political parties are particularly active in deploying artificial intelligence for election campaigns and spreading disinformation.

While Russia appears to have systematically incorporated AI-generated disinformation into its international strategy, other state actors from China and the Gulf states are increasingly targeting African nations to advance their narratives through sophisticated propaganda. The study also identified terrorist organizations, Islamist groups, and cyber criminals as significant users of AI-generated disinformation.

A key difference in the African context is the limited internet access in many regions. “In many African nations, access to the internet and social media networks is often prohibitively expensive, and in some areas not even available,” explains Allen. This technological gap has resulted in fewer sophisticated deepfakes and more reliance on simpler “cheap fakes” that require less technical expertise to produce.

The Democratic Republic of Congo provides a troubling case study of AI disinformation’s real-world impact. The ongoing conflict between Congolese government forces and the Rwanda-backed M23 rebel group has created fertile ground for disinformation. Allen notes that images and text content linked to Rwandan accounts have significantly influenced public opinion about the conflict.

“This reinforces the suspicion that certain political figures are exacerbating tensions between the two sides,” she said, adding that Rwanda has employed AI to flood social media platforms with content that drowns out opposing viewpoints—a tactic employed both in conflict zones and during elections.

South Africa, with approximately 26 million social media users, offered a prime target for AI-supported manipulation during its 2024 parliamentary election. The newly formed Umkhonto we Sizwe party, led by former President Jacob Zuma, circulated a deepfake video purporting to show US President Donald Trump endorsing the party. According to the researchers, this became the most widely shared AI-generated content during the election period.

In response, South Africa established the Real411 platform, allowing voters to report suspicious online political content, including potential AI-generated material. This represents part of a growing trend of fact-checking initiatives across the continent.

Burkina Faso presents another concerning example. Following the September 2022 coup that brought Captain Ibrahim Traore to power, videos featuring supposed “Pan-Africanists” urging citizens to support the military junta circulated widely on Facebook, WhatsApp, and other platforms.

“It turned out these were all avatars. They were not real people. They [had been] created to support a political narrative, in this particular case, supporting the coup,” Allen explained. The videos were created using the AI tool Synthesia, which she notes has also been employed by Chinese-based actors for similar disinformation campaigns.

The researchers suggest that African citizens can protect themselves by diversifying their news sources, though this has become more challenging as major platforms like X (formerly Twitter) and Facebook scale back their fact-checking operations, effectively shifting responsibility to users.

Despite lagging behind Europe in data protection regulations, Allen sees promising signs in the growing exchange of fact-checking practices across the continent and the development of platforms where users can report suspected disinformation. The study indicates that Africa’s evolving regulatory environment may actually benefit from learning from mistakes made elsewhere, potentially positioning the continent to develop more effective approaches to combating AI-driven disinformation in the future.