<![CDATA[Digital transformation in the financial sector encourages the application of Artificial Intelligence (AI) in the credit audit process. While AI is capable of improving the speed and accuracy of risk assessments, modern models such as deep learning are black box, raising issues of transparency and accountability—two things that are critical in credit audits that must comply with regulations and build stakeholder trust. This study uses the Systematic Literature Review (SLR) approach to synthesize the scientific literature that discusses the application of Explainable Artificial Intelligence (XAI) in digital credit audits. The SLR process includes: Query formulation (e.g. "explainable AI", "credit audit", "interpretability model"), Screening of studies based on relevance, methodological quality, and year range of publication, Extraction of key data (XAI method, dataset type, prediction model, tools used, and key findings), and Comparative synthesis analysis. Based on the Systematic Literature Review, it was found that the main XAI methods are SHAP and LIME, the most commonly used prediction models are Random Forest and XGBoost and several repeated weaknesses were found, including limitations in the representativeness of the dataset, the risk of overfitting, and the trade-off between the level of accuracy and the level of interpretability. Thus, the proper integration of XAI is expected to increase transparency, fairness, and trust in AI-based digital credit audits, while paving the way for more ethical audit practices and in line with regulatory demands.]]>
