2027: AI as INEC’s new headache

Before now, whenever issues of electoral malpractice were discussed, what readily came to mind included ballot-box snatching and stuffing, multiple registration, vote buying, result manipulation and underage voting, among others. But lately, technology has apparently introduced a new possibility. We now have what can be described as artificial voters!

How else do you describe supposed voters whose photographs on the voters’ register are not those of real human beings? This is something we did not have to worry about in previous election cycles.

This is not fiction. The Independent National Electoral Commission published its preliminary Register of Voters online on October 26, and Nigerians scrutinising it began circulating screenshots of suspicious-looking photographs, particularly entries associated with parts of Edo State. Concerns also emerged about alleged duplicate registrations, names of deceased persons and complaints by some genuine voters that their details were missing. INEC subsequently announced an investigation.

The commission’s ICT Director, Lawrence Bayode, has now disclosed that INEC’s Automated Biometric Identification System flagged 268 records involving suspected AI-generated photographs or other irregularities. It is important to note, however, that these are flagged records, not necessarily 268 confirmed fictitious voters.

The fact that irregularities appeared in a preliminary register doesn’t automatically mean they would have appeared in the final register. INEC says the entire purpose of displaying the preliminary register is to allow public scrutiny, claims, objections, corrections and the removal of irregular entries. The commission says the cleanup is underway and that the final register will not be certified until December 15.

We can therefore give INEC some credit because what this means is that the system is working to the extent that suspicious records were identified before certification. But another legitimate question is why those records were successfully created in the first place. While I agree that detection is good, prevention is better.

The point, therefore, is to establish how these suspicious-looking photographs got into the register. As far as I know, voter registration is procedural. It is not supposed to be like opening an anonymous social-media account. There are officials in charge, and biometric information is expected to be captured.

So, who created those questionable records? At which registration centres were they created? Was this deliberate manipulation or merely a technical or procedural weakness? Were officials involved, or was there a loophole in the system that was exploited?

AI enhancement of a photograph does not necessarily make the photograph fake. Many modern smartphones automatically use artificial intelligence to sharpen photographs, improve lighting, enhance faces and improve image quality.

INEC has therefore correctly said it won’t automatically remove genuine voters simply because AI was used to enhance their photographs. The real concern is with images generated artificially rather than genuine photographs merely enhanced by AI.

There should also be accountability on the part of INEC and its officials. Deleting questionable records is good, but it isn’t enough. If deliberate manipulation is established and carries no serious consequence, another person may try again. INEC should publish the broad findings of its investigation without compromising legitimate security details.

A clean register is the foundation of an election. Before ballot papers, voting, BVAS, electronic transmission and collation, there must first be genuine citizens entitled to vote. Those people are contained in the voters’ register.

Therefore, if the register contains duplicates, deceased persons or fictitious identities, or excludes genuine voters, the election begins with a credibility problem. The integrity of the voters’ register isn’t merely an administrative matter; it is fundamental to the credibility of the election itself.

This piece is not in any way anti-technology. Technology has helped Nigerian elections in the past. It has helped detect multiple registrations, biometric inconsistencies and other irregularities. BVAS has changed the accreditation process. Indeed, INEC says ABIS itself helped flag the 268 questionable records.

The lesson, therefore, is not that technology is dangerous. It is that every technological solution can create new vulnerabilities that must be anticipated. The same AI that can help detect fraud can also be deployed to create more sophisticated forms of fraud.

Related News

• Otti hails FG for providing funds for 1,000 residents
• 55,000 inmates awaiting trial in Nigeria — PPDC
• Ogun deputy gov, First Lady demand more appointments for women

The 2027 elections will arguably be Nigeria’s first major elections in the generative-AI era. By then, AI will affect far more than the voters’ register. Nigerians should prepare for fake photographs, deepfake videos, cloned voices, fabricated speeches, fake endorsements, false newspaper front pages, fabricated opinion polls and, ultimately, fake election results.

Imagine a convincing video appearing 24 hours before voting showing a presidential candidate supposedly withdrawing from the race. Or an audio recording supposedly revealing an INEC official plotting with a political party. Millions of Nigerians could see or hear such content before journalists or authorities have enough time to establish whether it is genuine. That is the larger danger.

The media has a huge responsibility. Newsrooms will face enormous pressure in 2027. A sensational video lands on WhatsApp, and everyone else seems to be publishing it. Social media is exploding with reactions. Do you publish first or verify first? Being first is useless if what you publish is false.

Media organisations need robust verification protocols before the campaign season reaches its peak. INEC’s electoral guidelines should also, where permitted by existing law, address deliberately deceptive synthetic media. Political competition cannot be allowed to become technological warfare without rules.

It is against this background that the PUNCH Media Foundation, the training arm of PUNCH Nigeria Limited, is organising a webinar themed, “AI, Misinformation and Nigeria’s 2027 Elections: Safeguarding the Truth, Protecting the Vote”. The webinar, scheduled for 11am on Wednesday, November 11, 2026, via Zoom, is expected to bring together stakeholders across Nigeria’s electoral, media, civil society and information ecosystem to examine the implications of AI-enabled misinformation for voter trust and electoral credibility. The conversation is timely because the challenge is no longer hypothetical. What has happened with the preliminary voters’ register should be treated as an early warning.

INEC needs an AI election strategy. The commission should publish, or at least clearly communicate, its strategy for dealing with AI-related threats ahead of 2027. It already has election-technology policies and should explain how emerging AI risks fit into its safeguards.

During the election season, INEC shouldn’t take hours to respond to viral falsehoods. Imagine a fake result sheet circulating online or a deepfake video supposedly showing the INEC chairman announcing the postponement of an election. INEC needs a clearly identifiable rapid-response channel through which citizens and journalists can immediately verify such claims.

Election security is traditionally understood in terms of police officers, soldiers, ballot boxes, polling units and collation centres. But in 2027, cybersecurity will also be election security. Protecting voter databases, INEC’s website, official accounts, result systems and communication networks will be as important as protecting physical electoral materials. A hacker doesn’t need to snatch a ballot box if he can undermine public confidence in the electoral system digitally.

INEC should also be careful not to label every irregularity an AI problem. Some problems may simply be the result of poor data entry, duplicate registration, human error, fraud, weak supervision or insider misconduct. Calling everything “AI-generated” could obscure more conventional failures that require equally serious attention.

The commission should not wait until election day to build trust. Electoral credibility isn’t created when the first ballot is cast. It is built months beforehand through transparent registration, credible candidate lists, clear rules, secure systems, open communication and prompt correction of errors. If Nigerians distrust the voters’ register before election day, every technical problem that occurs afterward will feed suspicion and conspiracy theories.

The 2027 elections will not be contested only at rallies, polling units and collation centres. They will also be contested on smartphones, databases and social-media platforms, and in the increasingly blurred space between what is real and what technology can manufacture.

INEC, the media and political parties still have time to prepare. Nigerians also have time to become more sceptical about what they see and hear online. The worst response would be to wait until election week before discovering what artificial intelligence can do to an election.

Before Nigerians cast real votes in 2027, INEC must make sure only real voters are waiting to cast them.