Swift, together with 13 global financial institutions, has conducted experiments, with results showcasing how AI and secure cross-border data collaboration could have on minimising fraud in global payments.
The experiments utilised privacy-enhancing technologies (PETs) to allow institutions to share fraud insights across borders securely. Among the participants, Swift mentions ANZ, BNY, and Intesa Sanpaolo. In addition to financial institutions, technology partners also participated in the experiments, with Google Cloud being one notable example.
Fraud defence across global financial operations
According to Swift, in one particular use case, the PETs enabled participants to verify intelligence on suspicious accounts in real time. This development could reduce the time taken to identify complex international financial crime networks and avoid fraudulent transactions from being conducted. In a different use case, participants leveraged a combination of PETs and federated learning, an AI model that visits each institution to train on its data, so that organisations can work together without sharing customer information, to identify abnormal transactions. The model, which was trained on synthetic data from ten million artificial transactions between all participants, was twice as effective in recognising instances of known fraud compared to a model trained on a single institution’s dataset.
After these successful experiments, Swift now plans to expand participation before rolling out a second phase of tests. These will utilise real transaction data and focus on demonstrating the technologies’ impact on real-world fraud. Representatives from Swift highlighted that, as the industry faces substantial losses each year to fraud, the company aims to enable secure sharing of intelligence across borders to support the reduction of this figure and enable fraud to be stopped more efficiently.
Before this, Swift has collaborated with the industry, leveraging modern technology to mitigate common issues and scale the speed, efficiency, and security of cross-border payments. The cooperative has been actively working on investigating the role of AI, with it currently having over 50 use cases across proof of concept, pilots, and live usage.
