News, Education and Events Decoding Digital Payments & Fraud

News, Education and Events Decoding Digital Payments & Fraud

How AI Is Taking on the Fraudsters - and Winning

How AI Is Taking on the Fraudsters – and Winning

Sponsored Content

[Editor’s Note: November is Machine Learning Month on CardNotPresent.com. This is the first article in a series from our sponsor exploring how, against increasingly sophisticated attacks, artificial intelligence and machine learning are being applied to online fraud prevention. Check back here throughout the month for more content that sheds light on the newest technology in fraud prevention and answers questions you might have on how it can impact your antifraud efforts.]

Some of today’s most advanced AI innovations are focused on solving the hardest problem in fintech: fraud. Predicting fraud at the speed of transactions means matching the pace of today’s fraudsters, and fraudsters change tactics constantly. In this high-stakes game of cat and mouse, the adversary is motivated, coordinated, and often brilliant.

In a recent report from Feedzai, an AI company that fights fraud at banks and merchants, the company’s Chief Science Officer Pedro Bizarro writes about the challenges of keeping up with the speed of fraud. Criminals are using the latest technology, for example, generating bot attacks that can add to carts 5 times faster than humans, often targeting newly launched or valuable products.

Fraudsters operating in this digital world count on hiding in piles of data and taking cover in seemingly innocent hiding spots. For example, at Feedzai we have discovered that null fields in device IDs are correlated with a 78 percent fraud rate, because many fraudsters use emulators to mask the names associated with their devices.

Concurrent with fraudster’s evolving techniques, we see customers demanding more and more immediacy, and payments are going increasingly real-time. In this adversarial, fragmented, and high-speed landscape, how can a business make good decisions, fast? Artificial intelligence carries an ironic promise: the most futuristic technology can bring the economy back to the way things used to be, during the age of horse and buggies, back when commerce was a one-to-one customer experience.

Business in the past felt secure and confident about transactions because they had complete understandings of the individuals behind them. Bankers and merchants lived in the same communities as their customers, and they engaged with each other in person.

Today, that trust has been diluted as commerce has gone digital, and as new and unknown fraud has permeated business with patterns that are often impossible to trace. For example, Feedzai data scientists have discovered high fraud rates associated with devices that were purchased within the past 24 hours.

Fighting fraud now requires transforming mountains of raw data into complete and accurate insights about transactions as they happen. And with its mission of turning big data into big insights, AI for fraud draws its lineage from theoretical physics. Physicists are trying to understand the guiding principles behind everything. To do that, some physicists work on breaking complex things down into their fundamental components, for example, by using particle accelerators. However, a second method available to physicists is to “go big,” by combining fundamental components to see what kinds of patterns emerge.

Andrew Tikovsky, the VP of Data Science at Feedzai, also holds a Ph.D in Physics. Describing emergent principles, he says, “Combine simple things and the new patterns that emerge will surprise you. At Feedzai, each feature we create is something simple. Whether a person did a transaction yesterday. Whether they’ve been in a physical place. These are simple things we know. And when we combine them together, we create a whole new concept of how they interact in a financial ecosystem.”

The result of this new data intelligence is that businesses can zoom in to hypergranular understandings about their customers. The old method of understanding behavior was to look at what was normal for a cohort. “Is this behavior normal for a gold card member? For a baby boomer?”

But businesses need information that’s richer than any cohort can provide. An organization needs to know: “Does this person, who’s a gold card holder and a baby boomer, typically spend more than $500 in a single transaction in a single shop? Does this person buy multiple items in one transaction?”

And as businesses seek to understand whether any given transaction is exhibiting “normal” behavior, AI lets decision-makers see in segments of one, rather than in segments of many. The result is that bankers and merchants can make tough decisions confidently in milliseconds, whether that means fighting fraud or serving a legitimate customer.

The emergence of the digital economy created a flood of data that businesses are only now beginning to make sense of, and as innovations in AI begin to keep pace with the evolving methods of today’s criminals, trust and knowledge are returning to one-to-one interactions once again.

Read the Next CNP Report Article