Traditional fraud detection methods struggle to keep pace with increasingly sophisticated criminal tactics, often relying on limited data from individual institutions. This limitation impedes the identification of complex fraud schemes that span multiple banks and jurisdictions.
In a bid to improve fraud detection, Swift is collaborating with Google Cloud to create anti-fraud technologies leveraging advanced AI and federated learning. The collaboration also includes technology partners Rhino Health and Capgemini, with Rhino Health responsible for developing the core federated learning platform and Capgemini managing the solution's implementation and integration.
The solution will utilise federated learning techniques and privacy-enhancing technologies (PETs), allowing financial institutions to collaborate without compromising sensitive data. This approach aims to improve the detection of cross-border fraud while preserving privacy.
By using federated learning, the project enables financial institutions to share fraud labels and enhance cross-border payment security, helping them stay ahead of fraudsters. The initiative reflects a growing effort to develop innovative technologies that support a safer and more resilient global financial ecosystem.
Traditional fraud detection systems generally rely on siloed data within a single financial institution, which limits the scope of analysis and the ability to detect emerging fraud patterns. A lot of financial institutions still rely on outdated fraud detection tools, such as rule-based systems, that fail to account for the growing complexity of fraud tactics. These legacy systems are often unable to correlate data across borders or multiple institutions, preventing them from identifying fraudulent activities that span different regions, platforms, and payment methods. As a result, fraud detection is reactive rather than proactive, allowing fraudsters to adapt faster than the systems designed to prevent them.
The financial industry is beginning to recognise that the traditional approach to fraud detection is no longer sufficient in the face of evolving threats. A UK Finance research revealed that more than 70% of financial executives believe that advanced AI and machine learning technologies are critical for improving fraud detection and prevention. These technologies can analyse large, complex datasets in real time and identify patterns that may go unnoticed by traditional systems. Additionally, AI can continuously learn from new data, improving its accuracy and effectiveness over time. This shift towards AI-driven solutions is necessary to combat increasingly sophisticated fraud schemes, especially in the realm of cross-border payments, where the scale and scope of fraud can often go undetected for longer periods.
Every day we send out a free e-mail with the most important headlines of the last 24 hours.
Subscribe now