ABSTRACT: Artificial intelligence (AI) and machine learning (ML) in fintech lending have revolutionized credit risk assessment, enhanced financial inclusion, but raised challenges to governance, such as algorithmic fairness, systemic stability, and regulatory sufficiency. The given paper provides a systematic literature review and analytical synthesis of AI-based credit risk assessment in the context of fintech lending, relying on 30 peer-reviewed articles that were published in 2012 and 2025. The paper compares the performance of ML models to traditional scoring principles, examines the dynamic of financial inclusion and digital exclusion, and suggests a five-stage governance framework called the Integrated AI Credit Risk Framework (IACRF), which is an original operationalization of the SAFE AI principles of statistical accuracy, algorithmic fairness, financial stability, and ethical governance. Results show that ensemble and hybrid explainable AI models have better classification compared to those without governance, and their use without good governance increases inequality, systemic risk, and exceeds regulatory capacity. The paper ends by giving specific policy suggestions to regulators, fintech companies, and multilateral development agencies to find a balance between credit innovation and financial stability as well as borrower equity.

KEYWORDS – Artificial Intelligence in Finance; Credit Risk Assessment; Fintech Lending; Machine Learning Credit Scoring; Financial Inclusion; Algorithmic Bias; Systemic Risk; Explainable AI (XAI)