Work

Systematic Literature Review on Ethical AI in Banking : Equity/Fairness in Credit Lending

Year: 2026
Type: review
Study design: Systematic Review
Abstract: Artificial intelligence (AI) and machine learning (ML) are increasingly used in credit scoring and loan decisions, promising faster and cheaper lending but also raising serious concerns about equity and fairness because models learn from past data that can contain discrimination. This thesis presents a systematic literature review of 30 peer reviewed studies published between 2020 and early 2026 that examine how AI based credit lending systems embed or reduce bias, how fairness is defined and measured, and how banks try to govern these systems in practice. Using PRISMA 2020 procedures and searches in Scopus, Web of Science, and EBSCO, the review maps temporal, geographic, and methodological trends, and summarises bias detection methods, fairness metrics, and mitigation techniques such as re-weighting, constrained optimisation and explainable AI. The study shows that most publications focus on consumer credit, mortgages and peer to peer lending in high income countries, that bias and disparate impact remain common, and that measured discrimination can often be reduced without a major loss of predictive accuracy, although no single fairness metric or mitigation approach works best in all contexts. Important gaps include limited evidence from emerging markets, weak links between technical solutions and day to day governance and audit processes in banks, and little knowledge about who gains or losses from algorithmic credit over the long term. Taken together, the findings give an overall picture of equity and fairness in AI‑driven credit lending and suggest both research and practical steps for designing credit models that are accurate and profitable but also more transparent, accountable and fair.
Language: English
Open Access status: green
OpenAlex ID: W7163725645
FWCI:
Cites: 53
Cited by:
Related to: 9