Mitigating Algorithmic Bias in AI-Driven Decision Systems for Financial Services

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Artificial intelligence (AI) and machine learning (ML) systems are increasingly embedded in financial services, powering credit scoring, lending, fraud detection, and risk management. While these technologies enhance efficiency and predictive performance, they also introduce algorithmic bias that can undermine fairness, regulatory compliance, and public trust. Despite rapid growth in fairness-aware machine learning research, limited clarity exists on how bias mitigation approaches are conceptualised and operationalised within financial contexts. This study addresses this gap through a systematic literature review of 32 scholarly and industry publications published between 2015 and 2025. Using NVivo-assisted thematic coding, word frequency analysis, and cross-keyword co-occurrence analysis, the study synthesises how algorithmic bias is defined, measured, mitigated, and governed in financial AI systems. Five dominant themes emerged: data nature and sources of bias, fairness metrics and evaluation, mitigation methods, ethical and regulatory governance, and organisational implementation. The findings reveal a strong concentration on data-centric and technical mitigation strategies, while ethical and governance dimensions remain comparatively under-integrated in operational discourse. In response, the paper proposes a socio-technical bias mitigation framework that conceptualises fairness as an emergent outcome of interactions between technical systems, organisational governance structures, and regulatory environments. The study contributes by bridging technical fairness research with organisational and regulatory implementation, offering structured guidance for responsible AI deployment in financial institutions.