Deepfakes, Fake Voices & Election Lies: Nigeria Races To Secure 2027 Polls
By TOSI ORE & MELVIN KOFFA
NIGERIA’S electoral authorities, technology companies and civil society groups are seeking stronger safeguards against artificial intelligence (AI) threats ahead of the 2027 general elections. They warn that deepfakes, cloned voices and coordinated disinformation could mislead voters and undermine confidence in election results.
The warning followed the AI and Democracy Forum 2026, organised by Yiaga Africa and partners in Abuja from 14 to 16 October. The three-day gathering brought together the Independent National Electoral Commission (INEC), the Centre for Journalism Innovation and Development (CJID), and Dataphyte Foundation, alongside technology companies and development partners.
Participants identified fabricated political announcements, manipulated images and targeted disinformation as serious threats to informed voter participation. Moreover, they cautioned that the spread of synthetic content could make it harder for the public to distinguish genuine evidence from falsehoods. Consequently, even authentic recordings or documents could be dismissed as AI-generated fabrications.
However, the forum acknowledged that AI could also improve electoral administration. It identified potential applications in election logistics, voter education, cybersecurity, accessibility and the detection of duplicate voter registrations or unusual result patterns.
Nevertheless, participants warned that errors in biometric matching or the extraction of figures from result sheets could disenfranchise voters and damage the credibility of election results. They therefore insisted that human officials must verify AI-generated alerts before taking decisions that affect voting rights or election outcomes.
INEC Chairman Joash Amupitan reinforced that position, declaring that algorithms must never determine the outcome of the 2027 elections. Instead, he said, AI should support human decision-making rather than replace it.
According to the forum’s outcome document, INEC committed to finalising an AI governance framework, adopting a binding human-oversight principle and deploying AI-powered cybersecurity measures to protect its systems. The commission also proposed an Electoral Information Integrity Coalition involving media organisations, editors, fact-checkers, civil society groups and telecommunications operators.
Furthermore, participants urged INEC to publish its governance framework and explain how officials would capture, electronically process, verify and collate physical result sheets. They also called for clear procedures to resolve discrepancies, independent pre-election audits of electoral technology, transparent procurement and accountability for technology vendors.
Meanwhile, Microsoft committed to uploading INEC’s final list of candidates for the 2027 elections into its systems and databases to help restrict the generation of misleading synthetic images of candidates.
Technology platforms also came under pressure to label AI-generated political content, disclose political advertising more transparently, improve moderation in Nigerian languages and respond swiftly to deceptive impersonation. Similarly, political parties and candidates were urged to adopt an ethical code governing their use of AI, while newsrooms were encouraged to establish digital forensics desks and educate the public about manipulated content.
Beyond Nigeria, the Economic Community of West African States (ECOWAS) and its Network of Electoral Commissions committed to finalising common guidelines for responsible AI use in West African elections.
Ultimately, the forum stressed that transparent systems, competent election officials, institutional preparedness and effective oversight would determine whether AI strengthens democracy or becomes a tool for electoral manipulation in 2027.
