Publication Details
Abstract
The speed of the digital financial ecosystem growth in the United States has driven the amount of sensitive financial information up, its speed, and its susceptibility. Given that cyber-threats, fraud activities, and mass data breaches are on the increase, financial institutions are turning to artificial intelligence (AI) to augment the security, confidentiality, and resiliency of transaction systems. This study explores how AI based solutions can be developed to secure sensitive financial data using Credit Card Fraud Detection Dataset 2023, a large scale, anonym zed dataset of more than 550,000 real-world financial transactions. This study examines how the high-level machine learning and deep learning algorithms can be used to detect fraudulent activities, abnormal transaction patterns, and enhance data protection systems in the U.S. financial systems. The performance of various AI models, such as the Logistic Regression, Random Forest, Gradient Boosting, and Deep Neural Networks are evaluated with a detailed methodology that includes the stages of data preprocessing, feature engineering, class-imbalance management, training models, hyper parameters estimation, and comparative analysis. The results have shown that AI-based fraud detection can be useful in ensuring the security of financial data as it can quickly identify possible fraudulent transactions and at the same time reduce false positives, which is of high importance in real-time payment systems. The study further explains how AI models could be executed in privacy-sensitive frameworks through the use of anonym zed features, which would ensure that the models comply with regulatory standards in the United States of America, including, but not limited to, GLBA, PCI-DSS, and FCRA. The subject of imbalanced datasets, emerging trends in fraud, and complex AI model interpretability is also described in the study. It ends by reiterating the fact that constant model adaptation, ethical use of AI, and use of privacy-enhancing technologies to reinforce financial data protection is essential. In general, this study provides useful information on the role of AI-based mechanisms in enhancing the safety of confidential financial information and achieving more robust financial systems in the U.S.