The rise of artificial intelligence is coinciding with a shift toward instant payments that are increasingly difficult to stop once fraud occurs. Real-time payments put a stopwatch on fraud prevention, leaving businesses with only moments to detect and respond to suspicious activity.
Striking the right balance between frictionless customer experiences and strong controls is becoming a critical challenge for businesses. In a recent PaymentsJournal Podcast, Dal Sahota, Global Director of Trusted Payments at LSEG Risk Intelligence, and Suzanne Sando, Lead Analyst of Fraud Management at Javelin Strategy & Research, discussed the importance of collaboration and highlighted how AI has become a double-edged sword—assisting fraud prevention teams while also giving criminals more sophisticated tools.
A Growing Concern
OpenAI tools have enabled scams to scale, increasing their ability to penetrate markets across the globe with minimal friction. Javelin’s research found that 88% of consumers are concerned that AI will be used to commit identity fraud against them.
“What I’ve been hearing more is voice can’t be trusted and video can’t be trusted,” said Sahota. “The scale has increased, meaning that the cost of committing fraud is very low, meaning that the potential gains that the frauds can go after are even more exponentially higher year on year.”
Sando added: “We’re all confident that the number one tool that’s going to be used by fraudsters is AI. We’re going to see a shift in focus to more manipulation and social engineering tactics versus just the more traditional way of trying to gain unauthorized entry into an account.”
Faster Payments, Faster Fraud
The rise of faster payments also means faster fraud. When money moves instantly from one domestic account to another, the sender often has little to no recourse to recover funds—regardless of whether the loss stems from fraud or simple error.
In cross-border payments, fraud exposure rises exponentially, and the likelihood of recovering funds is even lower. While some countries offer consumer and business protections that can partially offset these losses, reimbursement is typically limited to specific regulatory or legislative corridors.
Overall, the longstanding processing delays built into traditional payment channels have effectively disappeared. As a result, real-time detection and prevention of suspicious activity are no longer optional—they’re essential.
Detecting Legitimacy Is Paramount
Organizations should be analyzing every piece of data available to them to gain confidence in who is authorizing a payment or purchase. This includes the need for stronger shared network data and deeper network intelligence. Without access to that intelligence, organizations are likely to miss important signals—often at the exact moment they matter most. Detecting those signals in real time can prevent significant financial losses for customers and reduce future instances of identity fraud.
The challenge lies in navigating this process in real time: collecting and analyzing information using faster, more accurate data signals at speed. This requires evaluating biometric attributes tied to the device and the transaction, as well as determining what constitutes normal versus abnormal behavior.
How the Good Guys Use AI
More transactions are conducted digitally than ever before, with trillions of transactions and a quadrillion dollars in value exchanged each year. How is it possible to identify a bad or suspicious transaction amid all that activity? One emerging answer is the use of AI.
When combined with robust data and existing defense mechanisms, AI adds another layer of protection against attackers who are themselves using AI illegitimately. However, AI must play a proactive role—taking the offense in ways that can prevent fraud before it happens, not just detect it after the fact.
Criminals can take greater risks and move faster because they’re not constrained by AI governance or risk management teams. To keep pace, fraud prevention teams need strong collaboration and the elimination of organizational silos. This enables them to adopt AI responsibly as it evolves, close the gap with criminals, and ultimately get ahead of them.
Another major trend is the focus on authentication and identity proofing. Many banks are recognizing that they are losing confidence in the true identity of the user on the other end of a transaction.
“How can we trust that transaction if we can’t even trust the person who may or may not be authorizing it?” Sando said. “That’s going to be particularly important as we see a rise in deep fakes and synthetic identities that are aided by AI.”
Minimizing (but Not Eliminating) Friction
This is also an important moment for organizations to consider what their optimal level of friction should be. The conversation often centers on balancing friction with the consumer experience, but the goal should be less about eliminating friction entirely and more about applying it where it matters most. Effective friction comes from confidently verifying who is being paid or confirming that biometric data aligns with patterns observed across recent transactions.
Contextual signals such as biometric behavior, rich transaction data, and network and device intelligence provide valuable insight without creating unnecessary friction for consumers. These signals allow organizations to make confident decisions about whether fraud or suspicious activity is present without compromising the customer experience. When suspicious behavior is identified, authentication measures can then be appropriately escalated.
“When businesses make payments, typically to their suppliers, those can be 30, 60, even 90 days out,” Sahota said. “And one of the areas that we’ve been working on is how can we create tools to verify who they’re paying well in advance of when they pay. The friction is done much earlier, but it’s the right level of friction.”
Fostering Collaboration
True market leadership today depends on deep collaboration—partnerships that go beyond traditional boundaries to address challenges collectively. One area where this is starting to take shape is in the sharing of fraud insights across market participants, enabling faster detection and smarter prevention strategies.
“If we look at how our organizations manage fraud, whether that’s a bank, fintech or a multinational corporate, typically it’s done in some level of isolation,” said Sahota. “We need to get better with our cross industry and cross-border collaboration and data sharing. That’s where we have the strongest shot at reducing fraud and scam losses.”
But these efforts must evolve far more rapidly and on a larger scale. Fraud networks operate globally, and the response to them must match that scope and sophistication.
“A private-public sector collaboration and partnership would allow connections between everyone who has something to bring toward solving the problem,” Sahota said. “When we work together, we will get in front of the problem, and we will beat the fraudsters in their game that they play.”
