Google Cloud and Swift have joined forces to advance secure AI model training for financial institutions using federated learning techniques, a collaborative approach that maintains data privacy, the tech giant announced Tuesday (Dec. 10).
The initiative, outlined in a Google Cloud blog post, aims to develop a cutting-edge, privacy-preserving solution for the global financial ecosystem. Federated learning enables institutions to collaboratively train AI models without exposing sensitive data, a critical step toward enhancing security and confidentiality in the sector.
As part of the partnership, Swift will launch a sandbox in the first half of 2025, utilizing synthetic data in collaboration with Google Cloud and 12 international financial institutions. This sandbox will explore the viability of federated learning to combat fraud, improve cross-border payments, and enable real-time fraud label sharing across the financial network.
“This exploration will help the community validate whether federated learning technology can help financial institutions stay one step ahead of bad actors,” said Rachel Levi, Swift’s head of AI. “In turn, it will enable them to provide an enhanced cross-border payments experience to their customers.”
The project will also involve technology partners Rhino Health and Capgemini. Rhino Health is tasked with developing the core federated learning platform, while Capgemini will oversee the implementation and integration of the solution.
Key components of the initiative include:
- A federated learning server to manage collaboration among participants.
- Encrypted bank-specific data, which remains secure throughout the process.
- A global fraud detection model, leveraging aggregated insights to identify anomalies and combat fraud.
- A global anomaly detection model, trained to detect irregularities across financial networks.
Andrea Gallego, managing director for global go-to-market incubation at Google Cloud, highlighted the potential of the collaboration. “Our work with Swift exemplifies the transformative power of federated learning and confidential computing. By enabling secure collaboration and knowledge sharing without compromising data privacy, we are fostering a safer and more resilient financial ecosystem for everyone,” Gallego said.
This partnership builds on Swift’s ongoing efforts to integrate AI into financial services. In October, the organization announced the enhancement of its AI-driven fraud detection capabilities, set to roll out to the payments sector in January. By leveraging pseudonymized data from billions of transactions across its network, Swift aims to identify and respond to suspicious activity in real-time, further strengthening the global payments infrastructure.
Through these initiatives, Swift and Google Cloud reaffirm their commitment to leveraging AI to tackle industry-wide challenges and enhance the security and efficiency of financial systems worldwide.