Guinea's CENI Voter-Roll Overhaul — AI Age-Estimation and Biometric Deduplication Ahead of the 2020 Election

Guinea's Commission Électorale Nationale Indépendante (CENI) needed to consolidate 5.5 million newly enrolled voters with 6 million legacy voter records ahead of the October 2020 presidential election. To do this, it contracted the Slovak biometrics firm Innovatrics to build an Automated Biometric Identification System (ABIS) combining fingerprint and facial matching with a neural network trained for AI-based age estimation.

The project deployed 4,000 enrolment kits and trained more than 600 staff across 8 regional centres, delivering the full system within a six-month window ahead of the fixed election date.

According to Biometric Update, the age-estimation model flagged approximately 60,000 registrants who appeared to be under Guinea's minimum voting age for manual rechecking, while biometric deduplication separately identified roughly 170,000 duplicate, deceased or otherwise ineligible entries for removal. Independent Guinean media (Guinéenews), citing an electoral expert from the Organisation Internationale de la Francophonie (OIF), confirmed that the ABIS system's age-detection capability was a genuine addition to Guinea's electoral data infrastructure, though it did not independently verify the precise figures.

Guinea's October 2020 election was highly contested, marked by violence and opposition allegations of fraud; this record describes the technical AI deployment in the voter-roll process only, not a claim about the overall integrity of the election. The most detailed figures originate from an industry trade outlet reporting largely vendor-supplied data, so they should be read as vendor-sourced but plausible rather than independently audited.