Developing a Software Engineering Ethics Framework for AI Applications in Nigerian Public Services

The deployment of artificial intelligence in Nigerian public service applications, including social welfare targeting, police predictive systems, and revenue authority fraud detection, raises profound ethical questions that existing Western AI ethics frameworks address inadequately given Nigeria's specific postcolonial governance structures, ethnic diversity, and regulatory immaturity. This dissertation develops a software engineering ethics framework for AI applications in Nigerian public services, making original philosophical, theoretical, and applied contributions to AI ethics and software engineering ethics research. A critical interpretive methodology is employed across three phases. Phase one conducts a philosophical analysis of AI ethics frameworks including IEEE Ethically Aligned Design, the EU AI Act principles, and Ubuntu-based African ethics theories, identifying conceptual gaps for Nigerian public service AI contexts. Phase two conducts empirical grounding through 40 interviews with AI developers, public servants, civil society representatives, and affected citizens across six geopolitical zones concerning three live AI public service applications. Phase three synthesises an original Nigerian Public Service AI Ethics Framework (NPSAEF) incorporating six novel ethical principles: Ethnic Fairness Auditing, Accountability Chains for State AI, Communal Consent Architecture, Postcolonial Bias Testing, Civic Explainability Standards, and Reversibility Guarantees. NPSAEF is validated through a Delphi study involving 25 expert panellists and operationalised through a software engineering ethics checklist for government AI procurement. The dissertation advances African AI ethics as a distinct theoretical field within software engineering. Keywords: AI ethics, software engineering ethics, Nigerian public services, algorithmic fairness, postcolonial AI