AI’s Role in the Evolution of Credit Scoring in Banking

This thesis explores the pivotal role of Artificial Intelligence (AI) in transforming credit scoring practices within the banking sector. Historically, credit risk assessment relied on labor-intensive manual processes and traditional statistical models, which suffered from inherent limitations. The integration of AI models, encompassing machine learning techniques like Neural Networks and Random Forest, has ushered in a transformative era by adeptly handling non-linear relationships and efficiently processing vast quantities of dynamic financial data. This advancement significantly enhances risk management and financial inclusion by enabling the utilization of diverse, non-traditional data sources, including social media activity, e-commerce transaction histories, and behavioral patterns, to provide a more comprehensive and accurate borrower profile. AI-driven models offer distinct advantages over their traditional counterparts, particularly in accuracy, adaptability, and operational speed, by integrating a wider array of variables and continuously learning from new data inputs. Despite its transformative potential, the widespread implementation of AI in credit scoring faces significant challenges and ethical concerns. A primary issue is the potential for algorithmic bias, where AI models trained on historical data may inadvertently perpetuate discrimination in lending decisions. Another critical concern is the ”black box” nature of many advanced AI systems, referring to their opacity and complexity, which makes their internal workings difficult to understand or explain. To address these complexities, the emerging field of Explainable AI (XAI), with techniques such as SHAP and LIME, offers promising solutions by rendering complex models more transparent and interpretable. Concurrently, the regulatory landscape is rapidly evolving, with bodies like the European Central Bank (ECB), the Bank for International Settlements (BIS), and the International Monetary Fund (IMF) developing frameworks that balance innovation with risk control, model stability, and consumer protection