Nigeria’s 2027 general election is still more than a year away, but in Abuja on Monday the battle over its credibility had already begun. At the opening of the National Summit on Electoral Information Integrity, the Chairman of the Independent National Electoral Commission, Professor Joash Amupitan, issued a stark warning: artificial intelligence and foreign information manipulation could hollow out the vote long before citizens ever reach a polling unit. The summit, themed “Beyond the Ballot: AI, FIMI, and the Integrity of Nigeria’s 2027 Elections,” was supported by the European Union’s Support to Democratic Governance in Nigeria programme, and it brought together election officials, media organisations, fact-checkers, civil society groups and technology platforms. Amupitan described the weaponisation of information as the biggest technological challenge facing modern democracy, and he made the threat feel very real. Elections, he said, can now be undermined through synthetic media, automated disinformation, deepfakes, digital micro-targeting and Foreign Information Manipulation and Interference. These tools, once available mainly to well-resourced state actors, have become cheap and easy to use. Voice cloning, realistic fabricated video and coordinated bot networks can now be deployed by almost anyone with a smartphone, and that changes everything. Public awareness is no longer enough. In the past, a simple reminder to fact-check might have limited the damage; now, by the time a lie is caught, it has already been shared, believed and acted upon. The threat was urgent enough that INEC actually rescheduled the South-South leg of its zonal readiness assessment in Port Harcourt so that the commission’s leadership could attend the summit. With Nigeria’s presidential and National Assembly elections fixed for January 16, 2027, and state elections following in February, the official campaign season has begun, and parties and supporters are already using social media to promote candidates and denigrate opponents with false information. For Amupitan, the real battlefield is not the rally ground or the ballot box; it is the information space.
To make the danger clearer, Amupitan broke the electoral cycle into three moments where AI and FIMI could do the most damage. The first is the period before the election. Fake directives attributed to INEC, forged voter registration portals and false claims about updates to the voter roll can all discourage people from registering or convince them that their vote will not count. The second is the final stretch of the campaign, including the period immediately after campaigning ends. Under Section 98(1) of the Electoral Act, 2026, campaigns end 24 hours before polling day: January 14, 2027 for the presidential and National Assembly elections, and February 4, 2027 for governorship and state assembly polls. A deepfake audio or video released during that blackout is especially dangerous because the targeted candidate has almost no time to respond, and fact-checkers are still verifying the clip while it spreads through WhatsApp and Telegram. The third moment, and perhaps the most volatile, is collation. Fake announcements of results, or false claims that a vote has been cancelled, can be used deliberately to trigger unrest and undermine confidence in the outcome. Amupitan reminded participants of the 2023 Slovak parliamentary election, where an AI-generated audio clip purporting to show a party leader discussing vote-buying went viral during the pre-election silence. By the time fact-checkers, election authorities and media organisations could react, the damage was done. He also warned of what he called the “Liar’s Dividend”: the phenomenon where the mere existence of AI makes voters doubt even genuine material. “While technology evolves at the speed of light, the truth often travels at the speed of trust,” he said. That line captured the central dilemma of modern elections: even the best fact-checking may not keep pace with a well-timed fake.
Amupitan did not stop at describing the problem. He laid out the foundations of an INEC response, beginning with a shift in voter education from “how to vote” to “how to verify.” That means teaching citizens not only where to go on election day, but also how to test a suspicious message, how to check an official source, and how to avoid becoming an unwitting amplifier of falsehood. INEC plans to use “pre-bunking” through short audio messages in local languages and Pidgin on WhatsApp, meeting people on the very channels where disinformation actually spreads. The commission also intends to strengthen its digital infrastructure against FIMI, making official portals and channels much harder to clone or impersonate. Beyond prevention, there is a need for real-time response: mechanisms that can debunk viral misinformation within minutes rather than days, because a lie that circulates for six hours on WhatsApp can easily set the narrative for an entire election cycle. The commission is also pursuing formal platform partnerships with TikTok, Facebook, Instagram, WhatsApp, YouTube, Telegram, X and others, and an MOU is already in place with TikTok. But Amupitan was equally clear that INEC cannot do this alone. “INEC cannot police the truth alone,” he said. The responsibility extends to every institution that touches the public conversation. He called on social media platforms to create fast-track channels for removing demonstrably altered content, not through a formal complaint process that takes days, but quickly, while the damage can still be contained. He asked telecom providers and media houses to broadcast verified alerts without delay, and fact-checkers and journalists to be given direct access to INEC communication desks. Security agencies, he added, must learn to distinguish protected political expression from digital crimes designed to destabilise public order.
At the same time, Amupitan tried to draw a careful line between protection and censorship. Counter-disinformation efforts, he insisted, must never become a pretext for silencing criticism or controlling political debate. “The goal is not to police opinion, regulate political debate, or restrict citizens’ freedom of expression,” he said. The aim is narrower and more precise: to protect the authenticity of operational election information, so that voters can decide on facts rather than engineered deception. That distinction matters deeply in Nigeria, where social media is already a contested space and where parties and supporters routinely accuse each other of spreading lies. If the response to disinformation becomes a blunt instrument of state control, it will be both abused and ineffective. The way forward, in his view, is to build trust through transparency and speed. He urged the media to serve as the “first line of verification,” rather than simply repeating whatever enters their newsrooms. He invited civil society to help close the digital literacy gap in rural communities, where many voters rely on forwarded voice notes and messages and have few ways to test their authenticity. And he called on technology partners to develop detection tools suited to Nigerian languages and social contexts, not just English-language models trained on data from other countries. All of this should be designed with a central question in mind: how can institutions respond fast enough and credibly enough that citizens feel they can rely on official information? That is not an easy task in any democracy, and it is especially difficult in a country where trust in institutions is fragile, but it is the only way to prevent the information space from becoming a weapon.
Amupitan challenged the summit not to become “just another conference declaration.” He wanted concrete deliverables, and he listed three expected outcomes. The first is a set of inter-agency escalation protocols: a clear framework saying who detects a digital information threat, who verifies it, who escalates it, and who communicates it to the public during a digital information crisis. Without such protocols, even the best fact-checking effort can be undone by confusion over roles and delays in decision-making. The second is institutionalised fact-checking desks, with real-time links between INEC information officers, media monitors and certified fact-checkers. That would allow the commission and its partners to cross-check suspicious content quickly and to speak with one voice when something is confirmed as false. The third is simulation and stress-testing through tabletop exercises that model AI-driven election scenarios, from a deepfake of a returning officer to a fake INEC announcement cancelling polls in a major state. These exercises are not academic. They are a way to test response systems before they are needed under real pressure. Amupitan noted that INEC and a development partner recently conducted a strategic communication workshop that simulated some of these scenarios, and the lessons from that workshop helped shape the proposals now on the table. The ultimate goal, he said, is for Nigerian voters to be “empowered by information, not paralysed by it.” That phrase could stand as the mission statement of the whole summit. It is not enough to make disinformation difficult; citizens need to feel confident that they can find the truth, trust it and act on it.
The urgency of the moment was reinforced by Dr Lanre Arogundade, who spoke on behalf of the summit’s conveners. His message was direct: the battle over the 2027 elections is being fought on phones, not at polling units. With 75 days to the presidential and National Assembly elections, he said, awareness alone is insufficient against AI-driven disinformation. The statistics he offered were sobering. Of the claims reviewed by the Nigeria Fact-Checkers’ Coalition in 2023, 63 per cent were entirely false. Turnout in the 2023 presidential election fell to 26.7 per cent, the lowest since 1999. And 66.5 per cent of Nigerians now say they are deeply concerned about AI-generated political misinformation. Those numbers point to a dangerous spiral: false information undermines trust, distrust keeps people at home on election day, and low turnout makes the result easier to challenge and harder to defend. Arogundade and the conveners called for a trusted, coordinated and rights-respecting system that would allow institutions to anticipate threats, verify evidence, escalate incidents, communicate authoritatively and learn from each response. This is not simply a technical fix. It is an institutional culture change, from reactive scrambling to proactive preparedness. He urged the more than 300 participants to adopt an “Abuja Commitment” and a formal response protocol, while protecting free expression and media freedom. In other words, Nigeria must defend its democracy from digital manipulation without becoming less democratic in the process. The summit ended with a clear sense of the stakes: if 2023 taught the country how easily falsehood can shape an election, 2027 offers a chance to prove that institutions, media and citizens together can protect the truth. But the time to build that protection is now, not after the first convincing fake video goes viral.