Artificial Intelligence and the Future of Improving Credit Risk Assessment in Banking

Abstract

The paper reviews the use of AI and ML in improving credit-risk evaluation in financial institutions. With advances in AI, banks now apply machine learning, deep learning, and predictive analytics to analyse diverse datasets and generate more accurate credit-risk predictions. A structured questionnaire was administered to 300 purposively sampled banking and fintech professionals experienced in risk management. Perceptions of AI/ML adoption were analysed using descriptive statistics, correlation, regression, and ANOVA. Results show moderate agreement on benefits such as improved predictive accuracy, reduced Non- performing Assets, and better MSME credit assessment. However, barriers—including legacy systems, explainability requirements, data governance, and skill gaps—received higher concern. Correlation and regression indicate that AI/ML maturity strongly predicts better risk forecasting and lower NPAs, with no perceptual differences across institution types. The study concludes that AI/ML integration can significantly transform credit-risk assessment, but its full potential depends on strong data governance, a skilled workforce, and responsible AI practices. These findings offer practical guidance for banks advancing digital transformation of credit assessment models.