Artificial Intelligence Credit Scoring in Bank Financing: A Systematic Literature Review

Abstract

Advances in Artificial Intelligence (AI) have transformed the banking sector, particularly credit scoring systems, by improving the accuracy of credit risk assessment and the efficiency of financing decisions. However, AI implementation also presents challenges related to transparency, algorithmic bias, data protection, and governance. This study aims to analyze the development of Artificial Intelligence Credit Scoring in banking financing using a Systematic Literature Review (SLR) approach. The study follows the PRISMA 2020 guidelines and analyzes 30 scientific articles published between 2020 and 2025. The findings indicate that Random Forest is the most widely used algorithm, followed by XGBoost, Support Vector Machine (SVM), and Artificial Neural Network (ANN)/Deep Learning. The dominance of ensemble-based algorithms reflects the growing use of models capable of processing complex data patterns and improving credit risk prediction. Despite their potential to enhance operational efficiency and financing decisions, AI-based credit scoring systems face challenges involving the limited transparency of black-box models, algorithmic bias, personal data protection, and the need for comprehensive AI governance. As a conceptual contribution, this study proposes the Integrated Artificial Intelligence Credit Scoring Framework (IAICSF), which integrates the Data Foundation, AI Analytical Layer, Decision Layer, and Governance and Ethical Layer. In Islamic finance, the framework incorporates maqashid al-sharia, justice (‘adl), and customer rights protection as normative principles. The findings emphasize that successful AI credit scoring implementation depends not only on algorithmic accuracy but also on transparency, accountability, fairness, and effective governance.