From Perceived Gender Bias to Trust in AI-Enabled Digital Lending: A Procedural Fairness Framework for Nigeria

Abstract

Potential exists for AI-based and alternative data scoring to increase the availability of credit, and for discrimination based on gender or gendered data to further entrench structural disadvantage, or create an impression that gender data was used in the decision-making process. While research has previously considered statistical fairness, explainability, and technology acceptance, there is little understanding of the interpretive process through which discrimination allegations are turned into distrust. This conceptual article develops a procedural fairness process model for artificial intelligence enabled digital lending in Nigeria. It integrates contemporary research on procedural justice, algorithmic trust, and institutional assurance to outline a process in which individuals interpret decision clues, causally associate decisions with gender or gender-related data, evaluate procedural neutrality, correctness, accuracy, and appeal ability, and finally adjust their trust towards the algorithm and lending institution. Explanation, credible contestability, and successful human reassessment are considered as procedural fairness generators, rather than as alternatives to transparency. The perception of regulatory assurance dictates whether a fair process of platform lending leads to trust, while outcome favorability and attribution ambiguity determine initial assessments. The proposed framework separates perceived bias from measured disparity, procedural from distributive fairness, and trust from continued use despite financial necessity. Nigeria is viewed as a setting that defines the study due to the nature of informality, inequality of digital footprints, device sharing practices, different levels of literacy, and emerging consumer protection institutions. The article formulates revised hypotheses, competing theories, and a multi-method agenda for empirical research.