Financial institutions go to great lengths to shield their customers from fraud, but many are still utilizing an incident-driven approach. All too often, this strategy addresses a symptom while failing to uncover the fraud epidemic that spawned it.
One of the most alarming shifts in the evolution of fraud is how rapidly it has scaled beyond one-off scams. Financial crime is often perpetrated by organized syndicates who use ill-gotten funds to facilitate pernicious activities like human trafficking, drug trafficking, terrorism, and elder abuse.
Even worse, this trend is accelerating. Nasdaq Verafin’s 2026 Global Financial Crime Report found that illicit financial activity has surged by an estimated $1.3 trillion since 2023, representing a 19.2% two-year compound annual growth rate (CAGR).
This puts the scale of global financial crime at an estimated $4.4 trillion—a staggering statistic that still doesn’t capture many undetected or unreported fraudulent activities. What is equally alarming is that this surge is not necessarily the result of new, emerging fraud typologies.
“The fraud types haven’t changed much, and I don’t think they will,” said Greg Williamson, Head of Fraud Commercialization Strategy at Nasdaq Verafin. “We still experience the same type of attacks in card and the same type of ACH and wire attacks, account takeover—all of that is the same.
“What is changing rapidly is the scale and sophistication of the attacks,” he said. “The velocity has enabled the attackers to be much more effective in their attacks, they are able to execute their attacks at a higher rate, and be much more targeted in terms of knowing the consumer and their financial institutions.”
The Disappearance of Red Flags
The impact of this velocity is measurable. The Nasdaq Verafin report found that fraud, scams, and bank fraud schemes totaled an estimated $579.4 billion in losses globally last year, representing 9.2% CAGR since 2023.
One of the most important contributing factors is the rise of artificial intelligence. Authorized push payment (APP) fraud has long been a significant risk, but AI has supercharged this threat on many fronts.
For one, the tech allows bad actors to create convincing messages that are difficult to identify.
“In a traditional phishing attack in the past—to get your credentials or collect information from you—there were a lot of errors in that,” Williamson said. “Today, that can be pretty much fully clean, can look exactly like you would expect, can be created in a way that follows all the guidelines of what you should be doing for best practices. All those red flags that we ask consumers to look for are no longer in that process.”
The effectiveness of these attacks is fueling their deployment. Nasdaq Verafin found that 90% of surveyed financial professionals saw an increase in AI-driven attacks at their institution over the past two years. However, the improved messaging is only one driver behind this fraud trend.
Cybercriminals often couple these communications with social engineering techniques, where they attempt to pressure a user into making a payment. Consequently, this has fueled an uptick in romance scams, investment scams, business email compromise, and impersonation scams.
These scams are also more effective because they can be personalized, leveraging the massive troves of consumer data that are often readily available. Along with the information that has been filtered from the many data leaks and breaches in recent years, some consumers post significant amounts of data about themselves on social media.
The combination of all these factors have made fraud attacks difficult to discern, even for the most tech savvy consumers.
“Fraudsters don’t need to just send out a million emails or text messages anymore and hope to get a 1% response rate—which is already a pretty great result for them,” Williamson said. “Now they can send it to a couple thousand people that are very targeted and have very specific information about that person.
“When we look at how AI is impacting fraud, the sophistication isn’t just in a better written email, it’s in how that attack is targeted,” he said. “The schemes really don’t change much; they’re just better executed—and that’s a big difference. “
The Industrialization of Fraud
Much of this improved execution can be attributed to the rise of organized fraud rings. These syndicates can not only deploy all these tactics at scale, they can also quickly obfuscate funds via mule networks and instant payment rails—frequently before any red flags appear to the victim or institution.
Though these attacks can be devastating to institutions and individuals, they are increasingly only one aspect of organized financial crime. These fraud networks also perpetrate crimes such as drug trafficking and terrorism. They also commonly engage in human trafficking, which has become a sophisticated global epidemic conducted for financial gain.
“It’s an industrialization of the scam industry and the fraud industry,” Williamson said. “It’s better known now in terms of these massive organizations that are in some cases trafficking humans and forcing them into labor to do work and execute these scam calls, execute these checks, and engage with victims in a way where you need a large workforce to do that.
“Each person you add is a substantial increase in the amount of profit that the criminals ultimately make,” he said. “That’s been the business model, honestly, for a number of years.”
Although these ecosystems have become intertwined with fraud, institutions’ limited view of fraudulent activity means they can dismiss an attack as an isolated incident even when it is part of a global campaign.
Making matters worse, fraudsters are using AI to scale these activities and incorporating many of the same tactics legitimate organizations deploy, without the constraints of ethics or compliance.
“AI is going to change the game for those organizations as well,” Williamson said. “They can execute more attacks through AI, and they can move people the same way organizations think about it—into those more human-in-the-loop type interactions that are required in the more complicated schemes. When they get someone on the hook, then they can bring a human in to work on it.
“It’s scary when you think of the industrialization and the fact that these actually operate as enterprises,” he said. “The fraud organizations that we’re up against are well-funded, they’re well-connected with technology capabilities, and they’re able to move much faster in some cases in terms of their ability to make decisions and see what works.”
Protecting the Customer Downstream
This agility has enabled fraud organizations to rapidly shift gears in recent years. As financial institutions have invested heavily in building fraud detection and prevention systems, criminals have increasingly targeted the more vulnerable end user.
This shift has spurred the need for far-reaching intelligence. It is no longer enough to simply stop a criminal from gaining access to accounts or systems. Instead, institutions increasingly need insights into global money movement and changing behavioral trends.
Because these perspectives are outside many institutions’ scope, it has become clear that no single organization can combat fraud alone.
“Now more than ever, understanding where money is going is probably the most important thing,” Williamson said. “Sharing intelligence on the accounts and the fraud that you’re seeing, where those transactions are going to, and understanding the behavior of the receiving account becomes more critical. Understanding the threats that your peers are seeing becomes much more critical in terms of some of the capabilities and points of execution.”
To address this gap, there have been significant recent efforts to develop industry-spanning shared intelligence solutions.
For example, the surge in scams during the pandemic drove the creation of the Australian Financial Crimes Exchange. AFCX began as a collaboration among Australia’s leading banks but has rapidly grown to include telcos, digital assets exchanges, and government agencies.
This overarching model is crucial to root out fraud organizations—a task that should not be solely relegated to the financial services industry.
“It’s understanding where consumers are engaging, understanding what intelligence can be collected from a social media site or a telecom site, and linking that directly with the financial information and the banking information—because that’s where the scam and the fraud is occurring,” Williamson said.
“The more we understand about the devices, the type of conversations that are occurring, and the people who are engaging in certain activities on those platforms, the better we can protect that customer downstream.”
Bringing AI Into the Solution
While these solutions represent positive steps forward, they are just one step toward addressing the global financial crime epidemic.
Given the complexity and speed of today’s threats, there must be a concerted global response. Although cooperation is essential, any effective strategy should also incorporate technology, especially AI.
“We talk a lot about AI and the risk that it brings, but not always about the importance of bringing AI into the solution,” Williamson said. “It’s easy to say, ‘AI is going to make fraud or scam threats much more difficult to detect.’
“The good news is as fraud fighters we have access to AI as well, and this will enable us to fight emerging fraud and scam activity much mor effectively. AI will enable us to identify behavioral patterns and ingest new signals to better protect consumers. Going a step further, agentic AI will enable a whole new scalable workforce that can help institutions more effectively monitor accounts, triage risky activity quicker and enable quicker detection of fraud and scams.
These tools are critical to aid financial institutions that find themselves on the front lines of a challenging battle against fraud. However, the combination of technology, expertise, and collaboration can offer these organizations much needed tools and defenses.
What will best enable AI’s ability to mitigate fraud and scams and is the data it has access to. The data and insights available o AI models and agentic workers is what is going to differentiate their effectiveness to protect consumers and institutions. Williamson said. “That’s where the consortium approach is important—having visibility to understand not just the behavior at your one financial institutions , but the behaviors and insights across a network of institutions, pairing all of these insights together with a consortium view unlocks much greater AI capabilities. ”