AI regulation and democratic integrity in West Africa: Policy responses to data, platforms, and misinformation

Abstract

This paper examines artificial intelligence as a democratic governance problem in West Africa, with Nigeria as the principal case and Ghana, Senegal, Benin, and ECOWAS as regional comparators. Existing AI governance debates often treat data protection, platform regulation, electoral misinformation, digital identity, surveillance, and algorithmic bias as separate issues. This paper identifies separation as the central gap. It argues that, in weakly institutionalised democratic settings, AI operates not only as a tool for innovation or service delivery but also as a technology of classification, visibility, persuasion, exclusion, and public Power. The paper therefore develops a democratic data integrity framework built around five audit questions: whether data are lawful and representative; whether models are locally tested for bias and error; whether public institutions can explain and review automated decisions; whether affected citizens can obtain correction or remedy; and whether electoral and civic spaces remain open under platform Power and synthetic media. Using an analytical policy-review method, the paper compares official AI strategies, data protection laws, electoral reports, fact-checking evidence, regulatory actions, digital identity records, regional instruments, and recent African peer-reviewed research. It finds that the main challenge in West Africa is not the total absence of rules. Nigeria has adopted a national AI strategy, established a statutory data protection commission, used biometric electoral technologies, and created electoral AI capacity. Ghana has an AI strategy and data protection regime, while Senegal, Benin, ECOWAS, and the African Union provide wider regional reference points. The weakness lies in implementation, procurement transparency, audit capacity, surveillance oversight, platform accountability, and accessible remedies for ordinary citizens. The paper concludes that high-risk AI in West Africa should be treated as a form of public Power that must be lawful, contestable, auditable, and democratically accountable.