Artificial Intelligence as a Proactive Tool in Auditing Credit Risk: A Predictive Model for Iraqi Banks

Credit risk management has evolved significantly with technological advancements, particularly the application of artificial intelligence (AI), which offers enhanced decision-making tools for financial institutions. While AI has been successfully applied globally, particularly through machine learning models like XGBoost and SHAP, the use of AI in Iraqi banks remains underdeveloped due to infrastructure and regulatory challenges. Despite international research, there is a lack of practical, applied AI models specifically designed for the Iraqi banking sector, particularly those that incorporate local financial data and align with the country's unique context. This study aims to develop a predictive model using AI (XGBoost) to assess credit risk in Iraqi banks, evaluate the feasibility of AI models in this context, and propose solutions to overcome the challenges of adoption.  The developed model achieved high prediction accuracy (96%) and perfect precision (100%), demonstrating its effectiveness in classifying credit risk. SHAP analysis identified net profit, return on equity, and return on assets as the most significant variables. This study provides a tailored AI model for Iraqi banks, filling the gap in literature with a practical, applicable framework that can enhance risk management and decision-making in the Iraqi banking sector. The findings suggest that AI can significantly improve credit risk assessment in Iraqi banks, although the implementation of supportive infrastructure and regulatory frameworks is critical for broader adoption.