The global messaging system collaborated with 13 banks on experiments using privacy-enhancing technologies (PETs) to let institutions securely share fraud insights across borders, according to a Monday (Sept. 15) press release.
In one instance, the PETs allowed participants to verify intelligence on suspicious accounts in real time, “a development which could speed up the time taken to identify complex international financial crime networks and avoid fraudulent transactions being executed,” the release said.
In another case, participants employed a combination of PETs and federated learning, or an AI model that “visits” institutions to train on their data locally and lets them work together without sharing customer information, to spot anomalous transactions, per the release.
Trained using synthetic data from 10 million artificial transactions, the model was twice as effective in identifying fraud than a model trained using a single institution’s dataset, the release said.
“These experiments demonstrate the convening power of Swift as a trusted cooperative at the heart of global finance,” Rachel Levi, head of AI for Swift, said in the release. “A united, industry-wide fraud defense will always be stronger than one put up by a single institution acting alone. The industry loses billions [of dollars] 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.”
Advertisement: Scroll to Continue
In the wake of these experiments, Swift plans to widen participation before beginning a second round of tests, which will use real transaction data in hopes of demonstrating the technologies’ effect on real-world fraud, the release said.
When it comes to preserving trust in financial transactions, sharing data is important.
“It’s a team sport,” Entersekt Chief Product Officer Pradheep Sampath told PYMNTS in August. “And the thread that binds us all together is data that’s actionable, shared in good faith, and governed responsibly.”
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
