Harnessing AI in the fight against payments fraud

Artificial Intelligence (AI) is revolutionising how businesses operate. Faced with increasing fragmentation, new regulatory requirements and rapidly evolving external threats, financial service providers are turning to AI too. Not only to enhance their existing processes, but to deliver AI insights to elevate their offerings as well.

AI has immense potential to be a key enabler of instant and frictionless international transactions. And our latest collaborative innovation experiments aim to show how AI can play an important role in the fight against financial crime.

Leveraging AI to meet the global fraud challenge

This comes at a timely moment.

In 2023 alone, global fraud incurred a staggering cost of $485 billion, as reported by the Nasdaq-Verafin Global Financial Crime Report. It’s a growing challenge that’s becoming more acute for market participants with the advent of real-time payments and can incur significant reputational damage in addition to financial cost.

Our role as a trusted entity at the heart of the financial industry – combined with the breadth and reach of the transaction data flowing across our global network of over 11,500 institutions – makes us uniquely placed to unite the global community and tackle the systemic fraud challenge. Addressing this issue will also significantly help the industry achieve the G20’s goals of increasing the speed of cross-border payments.

Piloting a new AI enhancement for Payment Controls

Starting this month, we’re piloting an exciting new enhancement to our existing Payment Controls Service, which enables financial institutions to flag or block anomalous payments before they are made.

The pilot will involve five Payment Controls customers, including India’s Axis Bank, and test a new approach that uses AI-based algorithms to help them better detect fraud in transactions. The new AI model will be trained using historical patterns of activity on the Swift network to create a more nuanced and accurate picture of potentially fraudulent activity.

Importantly, this work will use live, in-production traffic data so that the findings of our pilot have real-world application. Working alongside pilot participants, we’ll test and refine the new enhancement with the ultimate aim of making it generally available to all 500+ Payment Controls customers worldwide.

Developing a pan-industry defence against fraud

There’s a growing consensus that global information sharing and collaboration between financial institutions is critical to the fight against fraud. Yet a recent report by the U.S. Department of the Treasury highlighted that current information sharing on fraud between financial firms is limited, when compared to the more common practice of sharing information about cybersecurity threats.

In this context, there has been a drive by countries and regulators to encourage initiatives that leverage recent advancements to ramp up anonymised intelligence sharing about fraud. But, while the number of projects in this space is growing, they typically lack global reach.

That’s why we’re launching another new initiative that builds on Swift’s trusted position at the heart of the financial industry to explore new opportunities for financial institutions to share insights and collaborate and improve fraud detection at the global level.

The initiative brings together an industry pilot group of 10 leading financial institutions across the world – including BNY Mellon, Deutsche Bank, DNB, HSBC, Intesa Sanpaolo and Standard Bank – to trial the use of recent technological breakthroughs to share intelligence and improve the financial community’s ability to detect fraud.

Enabling secure data collaboration

During the trials, we’ll test the use of both secure data collaboration and federated learning technologies. Initially, we’ll enable data collaboration by building a secure infrastructure to allow financial institutions to exchange relevant information with strong privacy-preserving controls. We’ll then apply federated learning to train our AI anomaly detection model on a much richer and broader dataset to identify patterns and gather insights. When combined, the techniques will enable us to arm our community with a defence against fraud that is far greater than any institution can achieve alone.

This work showcases the power of collaboration in creating solutions that raise the bar in the fight against fraud and greatly enhances our AI fraud detection model – all in a way that is wholly anonymous, secure and confidential.

The preparation phase of the pilot initiative kicked off earlier this year, and we’ll present our initial findings at Sibos in October.

Adhering to responsible AI principles – every step of the way

As a trusted cooperative at the heart of the global payments ecosystem, we recognise the transformative potential of AI. So as we unlock the power of this technology together with our community, we’re making sure that the responsible use of AI lies at the heart of our approach.

This means that as we roll out our pilot initiatives, we’re doing so in accordance with a strong governance framework – built in collaboration with the industry – to ensure that accuracy, explainability, fairness, auditability, security, and privacy are integral to every aspect of our AI applications.

As the regulatory landscape evolves around AI, we’re also ensuring that our work is aligned from the outset with emerging global standards, such as ISO 42001, the NIST AI Risk Management Framework and the EU AI Act.

In addition to helping us deliver transformative solutions for our community, we’re putting AI to work within Swift as well. We’re leveraging the technology ourselves to solve industry challenges, make us more productive and resilient as a company, and deliver optimum value for our members.

What participants are saying

Isabel Schmidt, Executive Platform Owner, Payments Enablement, BNY Mellon, Treasury Services, said: “Combatting fraud successfully will require industry collaboration, collective action, and strong leadership from operators like Swift. We’re excited to help lead the first wave of innovation in this space and look forward to seeing even more banks join the initiative.”

Ole Matthiessen, Global Head of Cash Management & Head of Corporate Bank APAC, Middle East & Africa at Deutsche Bank, said: “Our passion for holistic payment solutions is founded on expertise, technology, and innovation; not only in terms of our own strategic product development, but importantly through our strong dedication to drive partnership and collaboration across the market. Through these industry initiatives, we have the potential to enhance our ability to detect fraud and reduce operational costs, whilst delivering a premium customer experience and contributing towards a more secure market environment.”

Manish Kohli, Head of Global Payments Solutions at HSBC, said: “Building financial resiliency and strengthening risk management is a key priority for our clients and they are looking to us to help them better manage uncertainty and threat through our solutions while supporting their businesses to grow and transform. By utilising new technology and collaborating across the industry to fight fraud, we are helping protect our customers, our data and the bank.”

Enrico Canna, Head of Anti-Fraud & Customer Protection Center at Intesa Sanpaolo, said: “Extending and strengthening the collaboration of anti-fraud experts and the exchange of data also at an international level will improve the prevention of and fight against fraud.”