In a groundbreaking effort to enhance fraud detection in the financial industry, Swift is set to roll out a sandbox environment in early 2025, featuring synthetic data to prototype fraud detection systems based on historic payment patterns. Collaborating with 12 global financial institutions and leveraging Google Cloud as a strategic partner, this initiative builds upon Swift’s existing Payment Controls Service (PCS) and the success of a pilot program spanning Europe, North America, Asia, and the Middle East.
Federated Learning and Privacy-Enhanced Collaboration
At the heart of this initiative is federated learning, a cutting-edge approach to AI model training that maintains data privacy. Instead of centralizing sensitive financial data, federated learning trains models directly within financial institutions. Only the insights from this training are shared and aggregated via a central server managed by Swift, ensuring proprietary data remains secure.
Key partners in this initiative include Google Cloud contributing expertise in confidential computing and secure data collaboration. While Rhino Health is developing the federated learning platform. Capgemini is leading the implementation and integration of the system.
Industry-Wide Fraud Mitigation
Rachel Levi, Swift’s Head of Artificial Intelligence, highlighted the unique position of Swift as a trusted global network capable of fostering industry-wide cooperation. She emphasized that this initiative aims to validate federated learning’s potential to detect and prevent fraud through collaborative AI model training, without compromising data privacy.
Andrea Gallego, Managing Director at Google Cloud, echoed this sentiment, stating, “This partnership exemplifies the transformative potential of federated learning and confidential computing, fostering a safer financial ecosystem.”
Enhancing Cross-Border Payments
The initiative promises not only to combat fraud but also to improve the cross-border payment experience by enabling secure sharing of fraud patterns across institutions. By employing privacy-enhancing technologies (PETs) and federated learning, Swift and its partners aim to create a resilient and secure global financial landscape.
This collaborative effort marks a significant step toward leveraging advanced AI to protect financial institutions and their customers from evolving fraud threats.
Similar news:
SBS Partners with Red Hat to Drive Hybrid Cloud Transformation in Banking Sector
