|
Getting your Trinity Audio player ready...
|
Electoral and technology experts have called on the Independent National Electoral Commission (INEC) to strengthen its use of artificial intelligence to identify suspicious registrations and protect Nigeria’s voters’ register ahead of the 2027 general elections.
The experts said AI-powered tools could help the commission detect duplicate registrations, underage voter records, manipulated photographs and other irregularities that could undermine confidence in the electoral process.
Their recommendations come amid growing concerns over the discovery of suspected AI-generated photographs in Nigeria’s preliminary register of voters. INEC has confirmed that its biometric verification system flagged 268,000 records over suspected irregularities, including questionable images.
The concerns were highlighted in a report published by The PUNCH on Monday, November 9, 2026. The newspaper spoke with experts ahead of its forthcoming webinar on artificial intelligence, misinformation and the protection of Nigeria’s 2027 elections.
The debate has placed renewed attention on the integrity of the voters’ register, which is a critical foundation for credible elections. If records are duplicated, manipulated or improperly included, questions may arise about the accuracy of the number of eligible voters and the reliability of the electoral process.
INEC’s Director of Information and Communication Technology, Dr Lawrence Bayode, said the commission’s Automated Biometric Identification System (ABIS) could identify multiple registrations, underage records and other questionable entries.
He explained that the system flags suspicious records for further examination, while human adjudication determines which entries should be removed from the register.
Bayode said the commission was already reviewing the records identified by the system as part of its ongoing data-cleaning exercise. He also clarified that photographs genuinely belonging to voters would not automatically be rejected simply because they had been enhanced with artificial intelligence applications on mobile phones.
This distinction is important because AI-powered editing tools can improve the quality of genuine photographs, while generative AI can also create entirely artificial faces and images that do not represent real voters.
INEC’s position is that genuine voter photographs enhanced through ordinary phone applications should not be treated in the same way as fabricated images submitted as evidence of a person’s identity.Punch Newspapers+1
INEC investigates 268,000 flagged voter records
The discovery of suspicious photographs in the preliminary register triggered public concern about how such images passed through the registration process.
INEC previously disclosed that it had identified questionable photographs during its review of the preliminary register. The commission said the online version published on October 26 reflected the original preliminary register and did not include all the corrections and data-cleaning activities then underway.
The commission also confirmed that it was investigating how the photographs entered the register, including possible technical and procedural lapses. It promised to correct verified irregularities before certifying the final register, which was scheduled for December 15, 2026.
Bayode subsequently told The PUNCH that the biometric system had flagged 268,000 records. However, he did not disclose how many of those records had completed the review process.
The figure should not be interpreted as confirmation that 268,000 people were fake voters. It represents records flagged for suspected irregularities, which require further examination to establish whether they are invalid.
INEC also acknowledged that some registration officers had used their authorised login credentials to introduce AI-generated photographs and other irregularities into the register. The commission’s investigation is expected to establish how the breaches occurred and determine the appropriate corrective measures.
The development has increased pressure on the electoral body to explain its registration safeguards and demonstrate that questionable entries will not remain in the final database.Punch Newspapers+1
The issue is not limited to the technical accuracy of photographs. A credible register must also help prevent multiple registrations, underage enrolment and the inclusion of records that do not meet legal requirements.
AI can assist by comparing biometric information, identifying unusual patterns and directing suspicious entries to human reviewers. However, experts say the technology must operate alongside clear verification procedures and accountability measures.
The importance of human review is particularly relevant where a system flags a genuine voter because of a poor-quality photograph, a technical error or an unusual biometric record. Removing such entries without proper examination could disenfranchise eligible Nigerians.
Experts warn about AI-generated election misinformation
Beyond voter registration, experts have warned that artificial intelligence could be used to manipulate public opinion and discourage Nigerians from participating in the 2027 elections.
A fake notice announcing that voting has been postponed, for example, could prevent people from going to polling units. Similarly, fabricated videos or cloned voices of electoral officials could circulate false information about voting dates, registration procedures or election results.
The warning follows incidents during the June 20, 2026 Ekiti State governorship election, when AI-generated images circulated online as evidence of voter activity. Fact-checkers at the Centre for Democracy and Development determined that an image showing voters allegedly queuing to sell their votes was generated using artificial intelligence. Another image presented as evidence of massive voter turnout in Ado-Ekiti was also identified as AI-generated by the Nigerian Fact-checkers’ Coalition.
These incidents illustrate how fabricated images can be presented as genuine evidence during an election campaign. They also raise concerns about the speed at which misleading material can spread before fact-checkers and electoral authorities respond.
INEC Chairman, Professor Joash Ojo Amupitan, has defended the use of AI in election administration while warning about the dangers of synthetic media, including fabricated videos, manipulated photographs and cloned voices.
Speaking at the Beyond the Ballot National Summit on Electoral Information Integrity in Abuja on November 2, Amupitan said AI could support voter registration, logistics planning, data analysis and other administrative functions. However, he warned that the same technology could be exploited to spread false election announcements and undermine public confidence.The Sun Nigeria+1
Experts recommend stronger safeguards and rapid response systems
A Professor of Computer Science and Artificial Intelligence at Babcock University, Oludele Awodele, said INEC should adopt tools capable of identifying deepfakes and other forms of AI-generated misinformation.
He warned that individuals could use artificial intelligence to fabricate audio recordings, videos and images of political figures, electoral officials and public institutions.
Awodele recommended watermarking, deepfake detection systems and cooperation between INEC and social media platforms to identify and remove misleading content. He also called for sanctions against individuals who create, sponsor or distribute deceptive AI-generated materials.
The Acting Director of the Safeguarding Centre at the University of Lagos, Dr Abiola Akiyode-Afolabi, highlighted the risks facing women candidates and other vulnerable groups.
She said AI could be used to produce sexually explicit fake videos and other manipulated materials targeting women in politics. She recommended safeguards before, during and after elections to protect candidates and citizens from technology-enabled abuse.
Akiyode-Afolabi also pointed to the Violence Against Persons (Prohibition) Act 2015 as an existing legal framework that could offer some protection, although it does not specifically address artificial intelligence or technology-facilitated violence.
A certified information security professional, Precious Olives, recommended coordinated early-warning and rapid-response systems involving INEC, civil society organisations and security agencies.
He said natural language processing could help monitor public discussions, social media conversations, voter complaints, speeches and security concerns. Such systems could help authorities identify suspicious information and respond before it spreads widely.
The recommendations suggest that protecting the 2027 elections will require more than installing AI software. Electoral authorities will also need trained personnel, reliable verification procedures, cybersecurity protection and clear channels through which citizens can report suspicious information.Punch Newspapers
INEC also explores AI checks on election results
The commission’s interest in artificial intelligence extends beyond voter registration.
In October 2026, INEC disclosed plans to use AI to compare electronically entered election result figures with scanned copies of polling-unit result sheets before uploading them to its portal.
The commission’s Director of Information and Communication Technology, Dr Lawrence Bayode, said the system would flag discrepancies between the figures entered electronically and those appearing on scanned result sheets. Where the figures match, the system would allow the results to proceed for upload.
The initiative is intended to reduce human error and improve the accuracy of results published through the commission’s platforms.
INEC has also established an Artificial Intelligence Division within its ICT Department and is developing a framework to guide the responsible use of the technology.
However, the commission has cautioned developers against unauthorised access to election infrastructure and the use of sensitive electoral data without permission. It has also emphasised that technological innovation must not compromise security or public confidence.Vanguard News
The developments show that AI could become an important part of Nigeria’s election administration in 2027, from checking voter records to supporting result verification and providing citizens with reliable electoral information.
PUNCH schedules webinar on AI and the 2027 elections
The growing debate over artificial intelligence and election integrity has also prompted The PUNCH to organise a webinar titled, “AI, Misinformation and Nigeria’s 2027 Elections: Safeguarding the Truth, Protecting the Vote.”
The online event is scheduled for Wednesday, November 11, 2026, from 11am to 1pm, West Africa Time.
The webinar is expected to bring together representatives from the media, electoral institutions, security agencies, civil society and the technology sector to examine how Nigeria can benefit from AI while preventing its misuse during the election period.
The listed speakers include Inspector-General of Police Olatunji Disu, INEC National Commissioner and Chairman of its Information and Voter Education Committee, Mohammed Haruna, and YIAGA Africa’s Director of Programmes, Cynthia Mbamalu.
Other participants include Cable Newspaper Journalism Foundation’s Director of Projects, Mayowa Tijani, and founder of FactsMatterNG, Hannah Ajakaiye. The PUNCH Metro Editor, Joel Nwokeoma, is scheduled to moderate the discussion.
The webinar will examine the threats posed by AI-generated misinformation, the responsibilities of electoral stakeholders and practical measures for protecting voters from false information.
As preparations for the 2027 general elections continue, the central challenge for INEC will be ensuring that artificial intelligence strengthens the electoral process without creating new opportunities for manipulation or excluding eligible voters.
The commission’s review of the 268,000 flagged records, the investigation into suspicious photographs and the planned certification of the final voters’ register will therefore remain important tests of its commitment to electoral transparency.
