Nigeria's electoral commission used AI biometric voter accreditation (BVAS) for 93.4 million voters in the 2023 elections. Domestic observers found it worked properly in about 88% of sampled polling units; EU monitors recorded failures in roughly 10% of watched units.
93.4 million
Registered voters covered (2023 general elections)
88 %
Sampled polling units where BVAS functioned properly (Yiaga Africa)
9 %
Sampled polling units with BVAS malfunction and fix (Yiaga Africa)
25 of 240 (~10.4%)
Observed polling units with BVAS not functioning properly (EU EOM, presidential day)
Details
Maturity
Established
Promoter
Independent National Electoral Commission (INEC)
Period
2021-2023
Keywords
elections, public administration, biometrics
Context
Nigeria's Independent National Electoral Commission (INEC) introduced the Bimodal Voter Accreditation System (BVAS) -- a handheld device that matches a voter's fingerprint or face against the register -- as an upgrade to the earlier card-reader system, deploying it nationwide for the February-March 2023 general elections covering 93.4 million registered voters.
Activities
Independent domestic observation gives BVAS a broadly positive but incomplete record. Yiaga Africa's 'Watching the Vote' parallel vote tabulation, based on reports from 1,358 of 1,507 sampled polling units, found BVAS was used throughout accreditation in 99% of sampled units and 'functioned properly' in about 88% of them, while it malfunctioned and was fixed in 9% of units and malfunctioned and was replaced in a further 2%.
Results
The European Union Election Observation Mission recorded a higher failure rate on presidential election day, finding BVAS was not functioning properly in 25 of 240 observed polling units (about 10.4%), concentrated in specific states, though it reported markedly fewer problems in the state-level elections held three weeks later after INEC introduced corrective measures. The Commonwealth Observer Group likewise cited 'significant shortcomings,' attributing them to inadequate pre-election testing and limited polling-staff training, on top of separate failures in INEC's real-time results-upload portal (IReV) that further damaged public trust even where BVAS itself worked.
Conclusions
BVAS improved on Nigeria's previous accreditation technology and left an auditable biometric log that later fed post-election legal challenges, but neither Yiaga Africa's sample nor the EU mission's report translates into a single authoritative national failure rate, and no rigorous independent study isolates BVAS's net effect on fraud or turnout. The practice is best read as a large-scale, partially evidenced deployment with real, documented operational gaps rather than a proven success story.
Implementation
Indicative cost
High (€500k–€5M) — Nationwide handheld biometric device procurement and deployment; specific budget not itemised in sources.
Time to results
Medium (1–3 years) — Deployed for the February-March 2023 general elections; state-level elections followed three weeks later with corrective measures.
Staffing & skills
INEC (Independent National Electoral Commission), polling-unit staff nationwide