A Theoretical Framework for Ethical Artificial Intelligence Governance in the Nigerian Public Sector: Principles, Mechanisms, and Accountability Structures

Artificial intelligence deployment in Nigerian public sector decision-making raises profound ethical questions about accountability, fairness, transparency, and due process rights that existing AI ethics frameworks from Western regulatory contexts inadequately address. This study develops an original theoretical framework for ethical AI governance in the Nigerian public sector, grounded in Nigerian constitutional values, administrative law principles, and community accountability norms. Phase 1 conducted a systematic review of 42 global and regional AI ethics frameworks and governance models to identify their applicability and gaps relative to Nigerian public sector contexts. Phase 2 applied a qualitative Delphi study with 35 expert rounds involving AI researchers, public administrators, legal scholars, civil society representatives, and citizen focus groups across six geopolitical zones to derive governance principles and accountability mechanisms with democratic legitimacy in the Nigerian context. Phase 3 constructed the theoretical framework and subjected it to critical evaluation through adversarial expert review panels. Available AI governance theory from African contexts identifies colonial inheritances in universalised AI ethics discourse as conceptual limitations that an African-situated framework must explicitly address. The Decolonial AI Framework by Mohamed and colleagues, the Administrative Law Accountability Theory, and Ubuntu Communitarian Ethics provide the theoretical foundations. The original contribution is a four-level governance framework specifying principles, institutional accountability mechanisms, citizen remedy pathways, and inter-agency coordination structures specific to Nigeria's constitutional and administrative context. Keywords: AI governance, ethical AI, Nigerian public sector, accountability, algorithmic decision-making.