A major test run by Swift shows financial institutions can double their fraud detection power without sharing customer data, based on synthetic data trials.
Swift worked with 13 global financial institutions to test how artificial intelligence and privacy enhancing technologies could help detect and prevent cross border payment fraud more quickly.
The experiments used privacy enhancing technologies to let participants verify intelligence on suspicious accounts in real time, helping them identify complex international crime networks and avoid fraudulent transactions.
In a separate use case, these technologies were combined with federated learning, where the model trains locally at each institution so customer information stays private.
Trained on synthetic data from ten million test transactions, the model was twice as effective at identifying instances of known fraud as a model trained on a single institution’s dataset.
Swift plans to expand participation before a second phase using real transaction data to demonstrate the technologies’ impact on real world fraud.
The cooperative says secure collaboration at scale could help reduce industry fraud related costs.
Financial crime was estimated to have cost the industry US$485 billion in 2023.
Swift says it is exploring more than 50 AI use cases and earlier this year launched an AI enhanced Payments Controls Service to help small and medium sized financial institutions flag suspicious transactions in real time.
Participants included ANZ, BNY and Intesa Sanpaolo, alongside technology partners such as Google Cloud.
Rachel Levi, Head of AI at Swift, said,
“These experiments demonstrate the convening power of Swift as a trusted cooperative at the heart of global finance. A united, industry-wide fraud defence will always be stronger than one put up by a single institution acting alone.
The industry loses billions to fraud each year, but by enabling the secure sharing of intelligence across borders we’re paving the way for this figure to be significantly reduced, and allowing fraud to be stopped in a matter of minutes, not hours or days.”
Featured image: Edited by Fintech News Singapore, based on image by diloka107 via Freepik
