Swift and Google Cloud join forces to fight fraud with AI innovations

Swift, the global financial messaging service provider, collaborates with Google Cloud develop innovative anti-fraud technologies.

This initiative, developed in partnership with Rhino Health and Capgeminiaims to address the limitations of conventional fraud detection methods, which often struggle to keep pace with sophisticated criminal tactics that span multiple institutions and countries.

The partnership will leverage advanced AI and federated learning techniques combined with Privacy Enhancing Technologies (PETs) and Trusted Execution Environments (TEEs).

This will enable collaborative intelligence without compromising the privacy or security of sensitive financial data.

Federated learning allows participating institutions to train AI models on their own data and share only the learnings, not the data itself, with a central server managed by Swift.

This approach improves fraud detection capabilities while ensuring data confidentiality.

Fast plans to launch a sandbox environment in the first half of 2025, with Google Cloud as a strategic partner.

This sandbox will enable 12 global financial institutions to prototype learning from historical fraud data, using Swift’s anomaly detection model as a foundation.

The initiative builds on Swift’s existing payment controls service and follows a successful pilot program involving institutions in Europe, North America, Asia and the Middle East.

This collaborative approach offers several advantages. By sharing their information, institutions can identify complex fraud schemes that might go unnoticed by individual organizations.

Improved models lead to more accurate threat identification and fewer false alarms, reducing disruption to legitimate activities.

Knowledge sharing enables rapid response to new fraud trends, ensuring the system remains effective against evolving threats.

As more institutions participate, the efficiency of the system increases, creating a powerful network effect that strengthens fraud prevention for everyone involved.

The project emphasizes data privacy and security, using privacy-enhancing technologies and confidential computing techniques to protect sensitive information.

“This exploration will help the community validate whether federated learning technology can help financial institutions stay ahead of bad actors through the sharing of fraud labels, and thus enable them to deliver a improved cross-border payment to their customers. »

said Rachel Levi, head of artificial intelligence at Swift.

“Our collaboration with Swift illustrates the transformative potential of federated learning and confidential computing.

By enabling secure collaboration and knowledge sharing without compromising data privacy, we foster a safer and more resilient financial ecosystem for everyone.

said Andrea Gallego, Managing Director of Global GTM Incubation, Google Cloud.

Featured image credit: edited from Freepik