Artificial Intelligence and Electoral Democracy in Nigeria: Political Participation, Information Integrity and The Future of Democratic Governance

Review

Bisong and Suleiman examine AI as an emerging layer in Nigeria's electoral information environment rather than treating every digital election technology as artificial intelligence. That distinction is analytically useful. BVAS and IReV are positioned as digital electoral infrastructure, while generative systems, algorithmic curation, synthetic media, and AI-assisted verification create a different set of informational and institutional problems. The paper's central contribution is therefore not a claim that AI determined the 2023 election. It is a framework in which AI affordances affect information integrity, participation, electoral integrity, and democratic legitimacy through mediating conditions such as institutional capacity, platform governance, verification capacity, regulation, literacy, political trust, and inequality.

The documentary method is appropriately bounded. The authors triangulate legal and institutional material, peer-reviewed research, policy sources, and fact-checking evidence, and they explicitly refuse to infer voting effects from the circulation of manipulated or disputed content. Their contrast between a doctored political audio and another recording publicly described as a deepfake but assessed as not AI-generated makes the governance problem more precise: synthetic media creates both deception risk and authenticity-denial risk. The latter matters because a political actor can exploit the general possibility of fabrication to contest genuine evidence.

The paper moves from this information-integrity diagnosis to a distributed governance agenda involving INEC, legislatures, courts, platforms, civil society, fact-checkers, and citizens. It calls for provenance mechanisms, forensic capability, auditability, human oversight, platform transparency, evidence standards, and rights-preserving regulation. This is where the article becomes most relevant to governance practice, but also where its institutional design remains incomplete. The framework assigns responsibilities without specifying how conflicting authenticity determinations are resolved, what evidence threshold should govern high-stakes corrections, how provenance claims can themselves be contested, who enforces platform duties, or what redress follows when an authoritative verification is wrong. Provenance is treated as part of the solution, yet provenance can establish origin or chain of custody without settling legitimacy, context, or truth.

A reasonable defence is that the article is a documentary and conceptual synthesis, not a full implementation specification. That defence holds for the absence of detailed procedures, but it also defines the limit of the contribution. The paper identifies the institutional capabilities an electoral information ecosystem needs; it does not yet specify the decision rights, escalation paths, evidentiary standards, or correction machinery that would make those capabilities governable under disagreement. My assessment is that the paper succeeds at diagnosis and agenda-setting but stops short of a governance architecture. Naming INEC, courts, platforms, civil society, and fact-checkers does not settle which actor has authority when their authenticity determinations conflict. The next research step is therefore not simply more detection accuracy. It is testing institutional protocols for authenticity disputes, time-bounded correction, independent review, platform compliance, and public redress across Nigeria's multilingual and unevenly connected information environment.

Key Insight

The paper convincingly reframes electoral AI as an institutional evidence problem rather than a detection problem, but it offers a governance agenda rather than a governance architecture until authority for authenticity disputes, correction, platform enforcement, and redress is specified.