The Governing Void: Regulatory Gaps and Technical Imperatives in Nigerian AI-Driven credit scoring

AI-powered credit scoring has become central to financial inclusion in the Global South, utilizing alternative data to evaluate millions of unbanked consumers. However, regulatory frameworks designed to govern digital lending often address data privacy and legal non-discrimination duties without reaching the scoring algorithm itself. Examining Nigeria's dual regulatory landscape, namely the FCCPC's DEON Regulations (2025) and the Central Bank of Nigeria's (CBN) Open Banking framework, this paper demonstrates how a strict legal duty of non-discrimination and a sensitive data-access regime exist on either side of the credit algorithm, yet leave the model itself unexamined, unexplainable, and unaudited. We propose a legal-technical governance framework that bridges this gap by embedding formal algorithmic fairness metrics (Demographic Parity, Equal Opportunity, and Disparate Impact thresholds) and counterfactual explainability directly into regulatory reporting requirements under DEON, and we discuss the trade-offs, data limitations, and implementation barriers that any such framework must confront.