Generative AI-Based Financial Fraud Detection System

Financial fraud presents a serious threat to the stability and integrity of global financial systems. Traditional methods of detecting fraud often struggle to keep pace with fraudsters’ evolving tactics. In this study, a new method is presented to combating financial fraud by integrating various machine learning algorithms with generative artificial intelligence techniques. This work uses advanced deep learning techniques, such as convolutional and recurrent neural networks, to enhance financial system security. It focuses on improving fraud detection accuracy and reducing false positives and negatives by analyzing extensive datasets to detect subtle fraud indicators in real time. This initiative seeks to set a new standard in financial security, aiming to significantly reduce financial losses due to fraud. RNN has high accuracy (99.39%) but low precision, recall, and F1 score (40% each). CNN has slightly lower accuracy (99.29%) and relatively lower precision (33.33%) and F1 score (36.36%). FCNN shows high accuracy (99.49%) and precision (98.96%), with perfect recall (100%) and a very high F1 score (99.47%). It was observed that the GRU performs excellently across all metrics, with perfect scores (100%) in accuracy, precision, recall, and F1 score, indicating its strong ability to detect fraud while minimizing false positives and false negatives. This performance suggests that GRUs are highly effective for this type of task.