A Theoretical Framework for Governing Artificial Intelligence Ethics in Nigerian Public Administration: Institutional Perspectives

The deployment of artificial intelligence systems in Nigerian public administration raises fundamental governance questions that existing institutional frameworks are ill-equipped to address, particularly given the country's distinctive regulatory environment, administrative culture, and resource constraints. This study develops an original theoretical framework for governing artificial intelligence ethics in Nigerian public administration, grounding the inquiry in institutional theory, sociotechnical systems theory, and a comparative analysis of emerging AI governance models from Africa, South Asia, and the European Union. The research employs a multiple-case study design involving four federal agencies in Abuja that have adopted AI-assisted decision support tools in their operational processes, using document analysis, observation, and in-depth interviews with 38 policymakers, IT administrators, and civil society representatives. Empirical findings are synthesised to identify the structural, normative, and cognitive institutional pillars that must be present to enable ethical AI governance in a Nigerian public sector context. The original theoretical contribution, designated the Contextualised AI Ethics Governance Framework, integrates five governance dimensions: accountability architectures, algorithmic transparency obligations, bias auditing mechanisms, inclusive participation structures, and adaptive regulatory processes. The framework explicitly accounts for capacity constraints and colonial-institutional legacies shaping Nigerian administrative practice. The study advances AI ethics governance scholarship beyond its predominantly Western orientation and provides actionable guidance for the National Information Technology Development Agency's emerging AI policy agenda. Keywords: AI ethics governance, Nigerian public administration, institutional theory, algorithmic accountability, sociotechnical systems.