AI-Driven Models for Financial Fraud Mitigation: A Data-Centric Approach to Detecting and Preventing Fraudulent Transactions

The operation of new digital financial systems is greatly simplified by the incorporation of AI and blockchain networks. However, the techniques criminals employ have evolved, creating distinct challenges for conventional fraud detection systems. The contribution of this research is a framework for financial ecosystems that incorporates edge AI technology with blockchain for heightened security, alongside Generative Adversarial Networks (GANs) and Graph Neural Networks (GNNs) for decentralised fraud detection and response. The model is trained on heterogeneous financial datasets through GNNs with a multi-dimensional performance index assessment which showed exemplary gains in accuracy of detection, latency, and adaptability to changing fraud countermeasures. Moreover, credibility blockchains enhance system integrity by fortifying security measures against data breaches, while Explanatory Artificial Intelligence fulfils the regulatory necessity. The model’s design provides flexibility and adaptability to increasingly advanced requirements, reinforcing resilience against modern threats to financial infrastructures.