The incorporation of artificial intelligence (AI) in credit scoring and loan approval processes brings improved predictive accuracy compared to conventional methods while transforming financial decision-making. These black-box AI models make it difficult to understand their decision-making processes because they maintain internal operations that remain unknown. Consequently this creates concerns regarding transparency and both fairness and regulatory compliance. XAI stands as a vital instrument for financial responsibility because it delivers understandable explanations that resolve the problems with black-box AI systems. The research develops an XAI framework for credit scoring that evaluates five machine learning models including Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), and K-Nearest Neighbor(KNN) by using the Kaggle Loan Data Set that contains extensive applicant information. RF stands out as the top performer among the five models because it reaches 99% accuracy for loan outcome predictions. XAI techniques allow the model to demonstrate its decision-driving variables (credit history and debt levels) through feature importance analysis thus building trust and fairness. High performance of the model presents challenges because it runs complex calculations while also facing potential overfitting issues that require validation steps. The findings demonstrate the capability of XAI to supply transparent credit assessments with high accuracy that supports fair and compliant practices throughout the fintech industry and banks. XAI should gain more widespread use to establish an equilibrium between financial innovation and accountability within financial applications.