A Machine Learning Framework for Credit Risk Mitigation: Assessing the Impact of AI and Blockchain Integration

Credit risk remains one of the most critical challenges in modern financial systems, with rising borrower defaults, fragmented data infrastructures, and increasing regulatory demands threatening the stability of credit markets. Traditional credit risk models primarily reliant on statistical methods such as logistic regression are often limited in their ability to incorporate alternative data sources or adapt to non-linear borrower behavior. Furthermore, the integrity and authenticity of input data have come under scrutiny amid growing concerns about identity fraud, biased decision-making, and opaque lending processes. This study proposes a hybrid framework that integrates Machine Learning (ML), Artificial Intelligence (AI), and Blockchain technologies to improve the accuracy, transparency, and trustworthiness of credit risk assessments. Using a comparative research design, we evaluate the performance of standard ML models (e.g., Random Forest, XGBoost, Deep Neural Networks) with and without blockchain-based data verification and smart contract functionality. Public and simulated loan datasets were used to train and validate the models. Results indicate that the AI + Blockchain integrated framework significantly outperforms standalone ML models across key metrics such as AUC-ROC, F1-score, and fraud detection rate while also enhancing compliance through smart contracts and Decentralized Identity (DID) mechanisms.