Beyond E-Governance: Artificial Intelligence, Democratic Accountability, and the Governance Challenge in Nigeria

Abstract

Artificial intelligence (AI) is increasingly deployed across the Nigerian public administration system. However, dominant policy discourse continues to frame this deployment as an extension of e-governance which is considered a natural evolution of digital service delivery. This article challenges that framing by drawing on Schedler's (1999) two-dimensional conception of democratic accountability, answerability and enforcement. It argues that AI constitutes a qualitatively distinct governance challenge that existing e-governance frameworks are structurally ill-equipped to address. E-governance digitalised bureaucratic transactions whilst leaving decision-making authority with human officials. AI on the other hand reconfigures that authority, routing consequential public judgements through computational processes that existing oversight instruments cannot adequately reach. The empirical basis for this argument is original survey research involving 200 public servants and policy practitioners in Nigeria's Federal Capital Territory and Plateau State. Analysis reveals a statistically significant and practically meaningful gap between respondents' near-universal endorsement of responsible AI principles (aggregate mean = 4.46, SD = 0.39) and their assessment of actual accountability practices (aggregate mean = 4.00, SD = 0.63). The Wilcoxon signed-rank test confirms this divergence (Z = -9.23, p < .001, d = 0.77), whilst chi-square analysis identifies significant geographic variation between FCT and Plateau State respondents in practice perceptions (χ² = 18.37, df = 4, p < .01). These findings, interpreted through democratic accountability theory, suggest that Nigeria requires a new institutional architecture to govern AI responsibly) not merely a digitisation upgrade. The paper concludes with recommendations for regulatory, legislative, and administrative reform.