Introduction
On October 12, 2025, Cameroonians will vote in a key presidential election where incumbent Paul Biya is seeking another seven-year term in office.
According to the elections management body, Elections Cameroon (ELECAM), over 8 million eligible voters have 10 candidates from which to choose the next occupant of the Etoudi Palace, the presidential complex.
Campaigns have intensified nationwide ahead of the election. FactSpace West Africa has monitored the campaigns through social media accounts of the major candidates, their parties, the media, and the Cameroonian citizens.
Across social media platforms, videos and images of campaign activity are replete, with engagements centred around whether President Biya merits an extension of his term or Cameroon is ready for its third president since independence.
RelatedNews
The President’s video advert and the AI conversation trigger
On September 26, 2025, President Paul Biya, who has been in office since 1982, posted a 30-second video on his social media handles (X, Facebook, and Instagram). It wasn’t long before social media users hinted the video was AI-generated.
On X, the French version of the video garnered 1.6 million views with 476 comments, 1,600 reposts, and 701 likes. The English version had much lower interactions. On Facebook, only the French version of the video was uploaded, while both languages were uploaded on Instagram.
From the conversations FactSpace West Africa monitored, there were concerns that a President who has been in power for over four decades would resort to using AI to project his achievements; others were also concerned that there was no disclosure about the use of AI.
The President’s ad on social media, which caught the attention of the BBC, was therefore a trigger for the possible use of AI for the campaigns.
The SDF AI-images saga
The Social Democratic Front (SDF), one of the main parties in the election led by Joshua Osih, was caught in an AI-generated images scandal that started when a party official shared particular images of Osih’s campaign stop with journalists on October 30, 2025.
Fact-checker Annie Payep, who shared the images on X, captioned her post: “The SDF sent me images of what is supposed to be the Yagoua meeting on which the dates show October 30, 2025. I didn’t understand anything. I am looking closely, I am checking it out.”
Two of the four images shared were subjected to visual analysis and also run through Google’s SynthID detector tool. FactSpace West Africa concluded that the particular images were AI-generated.
- Both images had the four-tipped Google overview logo in the bottom right corner, the clearest evidence that the photos were AI-generated.
- In image 1, there is evidence that one person in front of the crowd was duplicated in the back, indicating tampering with the image.
- With respect to Image 2, even though Osih appears to be standing in front of the crowd, he is extraordinarily larger than those even closest to him.
Fighting AI with AI
The result from the Google SynthID detector showed that with respect to image 1, AI manipulation was detected in most parts, with small neutral parts.
In the case of image 2, however, the platform concluded that parts of the image appeared real, while other parts were doctored and neutral.
The SDF, in its October 4, 2025, statement signed by Henry Kejang, a senior official of the party, said they had seen the criminal use of AI to create and circulate images and videos of its candidate, stressing that the materials were neither from the SDF nor its candidate.
Fact-checker Payep, however, provided evidence to show that the said images were circulated to journalists on WhatsApp by SDF Campaign Director, Louis Marie Kakdeu, who later deleted them when controversy around the images started brewing online.
SDF’s belated AI-generated advert
While President Paul Biya launched his campaign with an AI-generated advert, the SDF opted to use AI in the later part of the campaign (archived here).
A video posted on the official Facebook page of its candidate Joshua Osih on October 9, 2025, had portions that were AI-generated per our analysis using the Google SynthID platform.
The Google SynthID in its results stated that the SDF video was “Made with Google AI (Video),” adding that “SynthID detected in all or part of the uploaded content.”
Further details showed that three parts of the video were generated using AI. The three parts, predominantly at the beginning of the video, account for about half of the 1:42-second video that was uploaded.
Main contender Issa Tchiroma accused – Fact-check
In the midst of election campaigns online, one of the frontrunners, who is leading a coalition, Issa Tchiroma Bakary’s team, was accused of sharing AI-generated images that sought to create the impression that he gathered massive crowds during his campaign stops.
“Even in Douala, AI is making a big impact! Yuck. Tchiroma is the only candidate drawing crowds that thanks to the AI that only geniuses know how to use on social media,” an X user captioned a post with three images claimed.
FactSpace West Africa ran checks on all three images through Google SynthID detector, and all returned the result that “Not made with Google AI. SynthID is not detected anywhere in the uploaded content.”
To add another layer of checks, we used the Hive Moderation tool for the three images as they appeared in the X post above.
| Image | One | Two | Three |
| AI score / Verdict | 100% genuine | 96.9% genuine | 88.7% genuine |
Additionally, a Cameroonian account on TikTok also shared videos that showed crowds similar to what was captured in the images. (here and here)
ACHPR on risks of AI and emerging technologies in the digital sphere
According to the African Commission on Human and People’s Rights (ACHPR) in its 2025’Study on human and people’s rights and artificial intelligence, robotics and other new emerging technologies in Africa,’ the commission underscored some risks presented by AI in the digital sphere.
“AI and emerging technologies pose significant risks in the digital sphere, notably in the areas of online hate speech, cybersecurity, and information disorder.
“The proliferation of disinformation, misinformation, and mal-information, including shallowfakes and deepfakes, creates a complex landscape of manipulated media that undermines trust in information sources.
“These technologies can spread unsolicited information, including non-consensual sexual images, which infringe on personal privacy and dignity. Further, AI, robotics and emerging technologies can be exploited to radicalize individuals and incite violence, amplifying societal divisions and threatening public safety.”
About Google’s SynthID tool
SynthID is a tool for watermarking and identifying AI-generated content generated by Google’s AI models.
These tools embed digital watermarks directly into AI-generated images, audio, text or video. For each modality, SynthID’s watermarking technique is imperceptible to humans but detectable for identification.
SynthID isn’t a silver bullet, but it is an important building block for developing more reliable AI identification tools and can help millions of people make informed decisions about how they interact with AI-generated content.
Researched by Alfa Shaban
Featured image is by courtesy of ELECAM handle on X
