Nigeria’s electoral commission says it has removed more than 1.6 million invalid voter records after biometric screening, with suspected AI-generated photographs among the irregularities identified. The cleanup puts a practical challenge of generative AI in focus: election systems must distinguish fabricated records from legitimate voters whose photos have simply been enhanced.
Watch Desk analysis
What happened
The Independent National Electoral Commission (INEC) says its Automated Biometric Identification System and further reviews have been used to screen the register ahead of the 2027 election. Of 268,000 records sent for further adjudication after the preliminary register was published, 217,722 were confirmed invalid and removed, while 50,278 were retained, according to Biometrics Update.
The commission’s account includes duplicate registrations, manipulated identity records and suspected AI-generated photographs. It says some registration officers used authorised login credentials to introduce AI-generated images and other questionable records. INEC says its investigation identified 54 personnel across 10 states as implicated in registration irregularities; it says they face administrative disciplinary action and possible prosecution. Those are the commission’s claims, not findings of guilt.
INEC says human adjudication is being used alongside biometric checks. It also distinguishes entirely AI-generated images from genuine voter photographs enhanced with AI-powered applications, and says voters should not be excluded solely because their photographs were enhanced.
Why it matters
The figures show both the reach and the limits of the cleanup. More than 1.6 million records have been removed, but that total covers invalid records generally, not a count of AI-generated photographs. The 268,000 figure is a separate group sent for further review, and the source says most of those records were ultimately removed.
For election administrators, detecting suspicious images is only part of the job. A system that flags records still needs a fair way to decide which should be removed, particularly when image enhancement can be legitimate. That makes the combination of biometric screening and human review consequential, rather than a technical footnote.
Our read
This is a substantial, concrete example of AI-generated imagery intersecting with election administration. INEC’s distinction between synthetic photographs and enhanced genuine images is important; automated suspicion should not become a shortcut to disenfranchisement. The useful test is whether the commission can explain how records are flagged, reviewed and challenged before the final register is certified.
What to watch
- Whether INEC publishes more detail on how its biometric system flags records and how people can challenge a decision.
- Whether the commission clarifies how many of the removed records involved suspected AI-generated photographs.
- Whether the final register is certified on the scheduled 15 December date.
Discussion spark: Should election authorities be allowed to use automated biometric screening to clean voter registers, or must every flagged record receive human review before removal?
Sources and evidence
- Source update (10 November 2026, 15:40 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.