The Jerusalem PostTrump: Iran exploring chemical, biological weapons, vows to 'wipe them out' if DC plot carried outPunchIG urged to curb election-related violence한겨레장동혁, ‘조작 징계’ 말 나오자 회의 끝내…“대여 투쟁 집중 공감대”InquirerAeta, farmers face displacement over Pax Silica hub – HRWZDF heuteAktuelle Pressemitteilungen des ZDFWirtualna PolskaChwila szczerości Trumpa ws. wojny w Ukrainie. "Zarabiamy mnóstwo pieniędzy"سكاي نيوز عربيةعقار تجريبي يستهدف سرطان البروستاتا المتقدم경향신문젠슨 황 없는 엔비디아 전사회의, 임시 CEO “혁신 향한 열정 변하지 않을 것”Daily MailAs zoo welcomes latest arrival who doesn't look like mum and dad yet... do you nose what species this seven-stone baby is?Free Malaysia TodayN. Korea’s Kim inspects work sites, university amid overseas travel speculationSoompi“Physical: Asia” Confirms Plans For Season 2Khaosod EnglishAnutin meets Putin in China, pushes Thailand–EAEU free trade deal
The Daily Newsstand · Free, Always
Thursday, November 19, 2026

AI, fake voter records raise new risks for 2027 elections

Translate

…Why media must not rush to publish election claims —Tech players

By Juliet Umeh

As Nigeria prepares for the 2027 general elections, voters may face a new kind of deception.

Nigerian Politics Coverage


The society is now rife with images, videos and audios that appear authentic but have been fabricated or altered.


The concern has grown after the Independent National Electoral Commission, INEC, said it removed 1,609,553 records from the voters’ register following a technical assessment. The review examined suspected multiple registrations, unusual registration patterns, substituted photographs and irregularities linked to registration transactions.


INEC also found cases in which some registration personnel deliberately bypassed established procedures, including provisions for exceptional circumstances involving persons with disabilities. The commission is investigating some flagged records.


The findings have prompted questions about how the irregularities occurred and whether synthetic images or other AI-generated material could be used to undermine confidence in the electoral process. There is, however, no evidence in the material available that AI-generated images caused the irregularities in the register.


Olaronke Alo of the BBC Disinformation Unit while speaking on Arise Television said its initial verification found that some of the images circulating online were present in the INEC register.


She said: “What we did was to do the initial verification. We went into the INEC portal and began to trace some of the names and pictures being posted on social media. We found that these images were on the INEC register.”


The unit also examined claims about inconsistent records, including one instance in which the same photograph appeared against different names.


Alo said it found examples of apparently incomplete or unusual names and records alleged to belong to deceased people, including former President Muhammadu Buhari.


She said the unit’s task was to verify claims circulating online, rather than assess whether INEC’s response was adequate.

False claims can travel fast


The risks extend beyond the voters’ register. AI tools can generate convincing images, videos and audio of people appearing to say things they never said. A fabricated statement attributed to a  political figure could spread rapidly before voters have time to verify it.


INEC has dismissed a social media poster and accompanying text claiming the commission had reaffirmed the leadership of the African Democratic Congress. The poster carried INEC’s logo and an image of National Commissioner Mohammed Haruna. With the matter before the courts, false claims styled as official announcements could mislead the public.


Other risks include political messages tailored to specific groups of voters, fabricated result sheets, screenshots purporting to show updates from INEC’s results portal, or cloned voices announcing false outcomes. Such material could create confusion and deepen doubts about the credibility of the process.

Political Analysis Reports


There is no evidence that these tactics have affected election results in Nigeria. But the possibility of manipulated content increases the work required of election officials, journalists and voters as the poll approaches.


Alo urged voters to examine the details of material they encounter online, asking who posted it, why it was posted and whether credible sources have confirmed it.


She cited a video appearing to show Burkina Faso leader Ibrahim Traoré addressing President Bola Tinubu in English. Viewers, she said, should consider whether the language matches what is known about the speaker and whether the video appeared through official channels.

She said: “If a president is addressing another president, it is expected that the national broadcaster of that country is picking up those broadcasts.”


Viewers can also look for inconsistencies in speech, mouth movements, skin tones and framing, she said. Reverse image searches may help establish whether a picture has appeared elsewhere and how it was originally used. People should also check official channels and note when platforms label content as AI-generated.


No single clue proves a post is fake. But checking the source and seeking confirmation before sharing can slow the spread of misleading claims.


Technology needs scrutiny


At a panel discussion on science and technology during election season, speakers said the Bimodal Voter Accreditation System, BVAS, and INEC’s Result Viewing Portal, IReV, could improve transparency. They cautioned, however, that technology alone could not guarantee credible elections.


Head of Research and Programmes at Kimpact Development Initiative, Oluwafemi John, said BVAS and IReV have different functions.


“BVAS helps with verification and accreditation,” he said, explaining that the device checks voters through fingerprints or facial recognition. After ballots are counted, the presiding officer uses BVAS to transmit an image of the polling-unit result sheet, Form EC8A, to IReV.


He noted: “IReV is not a collation point. We should not mix it up.”


He said questions remained about the legal and evidential status of some election technology records, and urged the public to distinguish between uploading results for viewing and formally collating them at designated centres.


A research analyst at the Centre for Journalism Innovation and Development’s Digital Technology, AI and Information Disorder Analysis Centre, Sanusi Sanusi, said online access to polling-unit results had enabled independent scrutiny.


He said: “Technology has really helped make election results transparent. If you know your polling unit, you can trace it on IReV and look at the result.”


But he raised concerns about BVAS devices’ age and performance, citing reports of accreditation difficulties and compatibility problems. He also questioned how differences between electronically generated figures and figures on physical forms were reconciled.


“If there is a disparity between the electronically generated figures and the physical figures, what is the reconciliation process?” he asked. “There is no guideline anywhere to explain that.”


Sanusi said inconsistencies in voter figures and reports of missing ballots needed investigation and clear explanations. He cautioned that discrepancies should not automatically be called evidence of rigging without further verification.
Verify before publishing


For journalists, the challenge is not only interpreting results but also assessing suspicious digital material before it reaches the public.


“Don’t run into publishing,” Sanusi said. “You need to develop your sense of doubt and figure out when something is worth investigating.” Tools that assess potentially AI-generated videos may offer clues, he said, but not definitive proof. “Always use multiple tools to verify.”

View the original on Vanguard →

KioskNews shows a cleaned-up reading view extracted from the publisher’s page — the original always lives on their site, not ours.