Advancing Financial Risk Management: AI-Powered Credit Risk Assessment through Financial Feature Analysis and Human-Centric Decision-Making.

The financial sector operates on a delicate balance of continuous cash flows between firms and customers, inherently exposing itself to credit risk. Effec- tive and precise credit risk assessment is paramount for maintaining finan- cial stability and mitigating potential losses. Leveraging the rapid advance- ments in artificial intelligence (AI) and machine learning (ML), this study proposes an automated, real-time credit risk evaluation framework. Utiliz- ing the World Bank Global Findex dataset, encompassing transactional data from 123 countries, the research focuses on extracting and analyzing financial features critical for risk prediction. A suite of machine learning algorithms, including Support Vector Machines (SVM), Logistic Regression, and Decision Trees, is implemented to model credit risk. These models achieve remarkable performance, with accuracy reaching 99% and an F1-score of 1.00, underscor- ing their robustness in identifying at-risk profiles. Interpretability is ensured using Explainable AI (XAI) tools, highlighting key financial features such as debit card usage, mobile banking adoption, and deposit patterns as the most influential predictors of credit risk. This study emphasizes the integration of domain expertise with ML models, advocating for a human-in-the-loop approach to ensure reliability and trust in automated systems. By focus- ing on model interpretability and transparency, the proposed solution aligns with regulatory requirements and enhances decision-making processes in the financial sector. This work highlights the transformative potential of AI in credit risk management, offering a scalable, data-driven framework for global financial institutions.