Skip to content

Enhanced privacy technology boosts swift fraud detection, doubling its efficiency in a testing phase

Banks, including Swift, carried out a trial of alleged privacy-protecting technology by processing ten million simulation transactions.

Enhanced privacy technology doubles the detection of real-time fraud during Swift's experiment
Enhanced privacy technology doubles the detection of real-time fraud during Swift's experiment

Enhanced privacy technology boosts swift fraud detection, doubling its efficiency in a testing phase

In a pioneering move, a collaborative, industry-wide initiative to tackle financial fraud has shown greater success than individual institution efforts. This innovative approach, spearheaded by Swift and 13 banks, could reshape the global financial fraud detection landscape.

The trials, involving financial institutions like ANZ, BNY, and Intesa Sanpaolo, along with tech partners such as Google Cloud, employed Privacy Enhancing Technology (PET) and federated learning. This novel method allowed institutions to exchange transaction data while preserving privacy and security, marking a significant stride in the battle against financial crime.

The model, trained on synthetic data from ten million artificial transactions, proved twice as effective in identifying known fraud cases compared to a model trained on data from a single institution. The use of federated learning, an AI model that trains locally with data from each institution without sharing customer information, was pivotal to this achievement.

The success of these trials suggests the potential for widespread adoption of Privacy Enhancing Technology in global financial fraud detection. The technology, when combined with federated learning, enabled participants to collaborate without sharing customer details, presenting a promising solution for the industry.

The trials demonstrate Swift as a trusted cooperation platform in global finance. The use of PET and federated learning facilitated secure cross-border fraud information sharing, enhancing fraud detection effectiveness using synthetic and local data.

The test resulted in a doubling of real-time fraud detection cases, underscoring the potential benefits of collective industry efforts in combating financial crime. If widely adopted, this technology could transform global financial fraud detection, fostering a more secure and efficient financial system for all.

Read also:

Latest