In May 2026, Nigeria experienced a revealing test of its preparedness for the age of synthetic media. A viral audio recording, apparently featuring President Bola Tinubu, circulated widely on social media. The voice made politically sensitive remarks about opposition parties, insecurity, public borrowing and the 2027 elections. The presidency condemned the recording, and presidential adviser Bayo Onanuga called for the prosecution of social media commentator Martins Otse, popularly known as VeryDarkMan, whom he accused of disseminating it.
The next day, FactCheckHub established that Mr Otse had not posted the manipulated recording. Someone had taken footage from one of his Instagram videos, added AI-generated audio resembling the president’s voice, and circulated the altered version. About a week later, police arrested Ifechukwu Dennis, whom the presidency identified as the originator of the fake voice. By then, however, the controversy had already drawn national attention, and the presidency had publicly sought the prosecution of a man who was not its source.
|
|
The episode exposes a problem that goes beyond the existence of a deepfake. The presidency was at once the injured party, the accuser and the first responder, and it named a culprit before anyone had established where the audio came from. Convincing synthetic content can now be produced far faster than it can be authenticated, and in that gap, institutions act on what they have not verified, journalists report claims of unclear origin, and the public forms opinions on fabricated evidence. With Nigeria preparing for another general election in 2027, the question is not only whether our institutions can identify fabricated material, but whether verification can remain independent of those with a stake in the outcome.
The problem is already global
Political figures elsewhere are already responding to the threat. On 5 October, Italian Prime Minister Giorgia Meloni applied to the European Union Intellectual Property Office to trademark her voice amid concerns about AI-generated impersonation. The application followed manipulated images of her that circulated online as authentic in May 2026, raising questions about how public figures can protect their identities when their voices and likenesses can be reproduced so easily.
|
|
|
|---|
The consequences can be serious during elections. In October 2025, days before Ireland’s presidential election, an AI-generated video circulated, falsely showing candidate Catherine Connolly announcing that she was withdrawing from the race. The video mimicked an RTÉ news report, Ireland’s national broadcaster, giving the fabricated announcement the appearance of legitimate news. Ms Connolly had made no such announcement and remained in the race, but the incident showed how synthetic media can create confusion at a critical moment, when voters, journalists and electoral authorities have little time to establish what is authentic. Nigeria faces the same risk in 2027.
Fake AI political content is already circulating in Nigeria
In September 2026, DUBAWA investigated AI-generated newspaper images that falsely attributed rent-policy promises to Atiku Abubakar. The fabricated pages claimed that he had promised Nigerians would no longer pay house rent if he became president, and that landlords who continued collecting rent could face imprisonment. The images were designed to resemble newspaper reports, borrowing the credibility of established news organisations to make the claims appear authentic. On social media, where such images are often shared without links to the original publications, readers could easily encounter the fabricated headlines without knowing they were false.
In July 2026, social media users accused Peter Obi of fabricating photographs to suggest that he had attended a European Council on Foreign Relations roundtable in Berlin. FactCheckHub subsequently confirmed that Mr Obi had attended the event, although some of the photographs he shared may have been digitally enhanced. The controversy quickly moved beyond the authenticity of the photographs to whether Mr Obi had attended the event at all. Questions about the images had become grounds for disputing an event that had actually taken place.
Stay Ahead with Premium Times
Follow us on Google News and never miss breaking stories, investigations, and in-depth reporting.
Add as a preferred source on Google
In December 2025, British-Nigerian boxer Anthony Joshua was injured in a car crash on the Lagos–Ibadan Expressway. President Tinubu publicly expressed concern, but photographs that circulated soon afterwards showing him at Joshua’s hospital bedside were AI-generated. By inserting the president into a real news event, the images created a political encounter that never happened.
The growing cost of verifying AI fakes
Someone with widely available AI tools can produce several versions of a fabricated political statement within minutes and spread them across multiple platforms. Establishing whether that material is authentic takes far longer. It means tracing the original source, examining the content for signs of manipulation, checking it against existing records and contacting the people involved, sometimes with specialist tools or outside help. Even after a claim is shown to be false, the correction rarely reaches everyone who saw the original. During elections, that gap can shape public opinion before journalists or electoral authorities have established the facts. The capacity to produce manipulated content is expanding quickly, while the expertise, infrastructure and cooperation needed to verify it are developing far more slowly.
The economics of synthetic media
Not everyone producing AI-generated material is trying to influence political opinion. Platforms such as TikTok, Facebook, YouTube, Instagram and X reward attention, engagement and audience growth, and their monetisation arrangements give creators a direct financial interest in content that travels. AI tools allow such content to be produced in large volumes at little cost. Fabricated celebrity endorsements, synthetic interviews, fake breaking-news clips and manipulated political statements may be made for money, political influence, publicity or fraud, but they draw on the same accessible tools and circulate through the same channels. For newsrooms and fact-checkers already working with limited resources, each increase in that volume adds to a verification burden that is growing faster than their capacity to carry it.
Is Nigeria prepared for 2027?
Elections compress the time available to respond. Voting happens within a fixed window, results are announced under intense scrutiny, and political developments provoke immediate reactions across the country. A fabricated concession speech, an AI-generated recording containing inflammatory ethnic or religious remarks, a false announcement attributed to INEC, or manipulated footage of electoral violence could each spread widely before journalists or security agencies establish what happened.
The reverse problem is just as serious. As deepfakes become familiar, politicians may find it easier to dismiss genuine recordings as fabricated, a pattern legal scholars Robert Chesney and Danielle Citron call the liar’s dividend. The Obi controversy showed how quickly doubts about images can become doubts about whether an event took place at all. Establishing that a recording is authentic may become as important as exposing one that is not.
Nigeria has fact-checkers, newsrooms, regulators and law enforcement agencies working on false information, but their procedures were not designed for the speed at which synthetic media circulates. A fact-checker may establish that a video is fabricated without having a direct channel to the platform hosting it. A party may deny a recording without offering evidence. And, as the VDM case showed, authorities may pursue someone associated with a viral clip before identifying where it came from.
Closing these gaps requires preparation before the campaign begins. INEC should set out clear procedures for responding to fabricated announcements and manipulated statements attributed to its officials. Parties and candidates should maintain known channels through which journalists and the public can confirm what they have said. Newsrooms should require verification before disputed audio, video or images are presented as authentic, and should state any remaining uncertainty plainly when the material is newsworthy. Fact-checkers, especially smaller ones, need forensic tools and training in source tracing and digital investigation. Platforms should provide escalation channels for credible fact-checkers and electoral authorities, explain how their synthetic-media policies will apply during the election, and respond promptly when content is shown to be deceptive.
None of this works if these institutions meet each other for the first time on election day. If a fabricated INEC announcement begins circulating at midday, electoral officials should already know whom to call for technical help. Journalists should know where to get confirmation, and platforms should have a channel ready for credible reports. Independent verification must remain possible throughout, particularly where political authorities are themselves parties to the dispute.
Two further points deserve attention. First, many of the AI models and detection tools used in Africa are built elsewhere, with uneven support for African languages and local contexts. Reliable tools and platform data are also often too expensive for smaller organisations. Collaboration among African newsrooms, universities, fact-checkers and public-interest technologists could ease this through shared expertise, regional investigations and verification resources suited to local information environments. Second, public education matters, but most citizens cannot authenticate sophisticated synthetic media on their own. They depend on journalists and credible institutions to do it for them, which is why sustained investment in investigative reporting and verification capacity is not optional.
The challenge before 2027
Nigeria has lived with propaganda and fabricated claims for decades. What artificial intelligence changes is how cheaply convincing material can be produced and how far it can travel. The cases already documented in Nigeria, alongside experiences elsewhere, are reason enough to strengthen the systems for identifying and responding to synthetic political content before the next major controversy rather than after it. As Nigeria prepares for 2027, the question is whether our institutions will be ready to establish what is true when it matters most.
Discover more from Premium Times Nigeria
Subscribe to get the latest posts sent to your email.
More News
