Towards Understanding Fairness Adequacy Testing in Financial Machine Learning Systems

Machine learning (ML) systems in finance raise concerns about fairness, bias, and regulatory compliance, especially in high-stakes areas like creditworthiness, lending, and risk assessment. Bias in ML can lead to systemic economic exclusion, disproportionately affecting marginalized communities. This paper explores fairness adequacy testing in financial ML models, addressing regulatory, technical, and ethical challenges while proposing bias mitigation strategies. Financial ML models face unique constraints, including strict privacy regulations such as the General Data Protection Regulation (GDPR), the Equal Credit Opportunity Act (ECOA) in the U.S., and Basel Committee fairness standards, along with reliance on proprietary data. While this work provides an overview of fairness challenges, the detailed scoping review will be conducted in a subsequent study to further analyze regulatory frameworks and industry practices. This paper aims to highlight key gaps and considerations that must be addressed to balance fairness and accuracy in ML-driven financial decision-making.