sponsored

Outsmarting Fraudsters with AI and Biometrics

biometrics1000x620
Dell EMC

In the world of payment cards and ecommerce, artificial intelligence and biometrics are now among the industry’s most potent weapons for fighting fraud. That was a key takeaway message from a recent presentation by Nick Curcuru, Mastercard’s vice president for global big data consulting.

In his presentation at the Dell Technologies’ “Unlock the Power of Data” virtual event, Curcuru offered an inside look at how Mastercard uses AI, machine learning technologies and biometrics to make card transactions seamless and frictionless for users, with sophisticated security from the edge to the core and back.

For starters, let’s consider the scale that Mastercard operates at. It’s mindboggling. In round numbers, the company has 2 billion cards in use in more than 210 countries and territories. It processes 165 million transactions per hour, using machine-learning algorithms and applying 1.9 million rules to examine each transaction. And it all happens almost instantaneously — in a matter of nanoseconds, Curcuru says.

With all of these transactions, the algorithms examine things like the cardholder’s buying habits, geographic location and travel patterns, along with real-time data on card usage. What are they trying to buy? Where are they are trying to buy it? What else have they bought in the same day? Each transaction is analyzed in terms of the rules that relate to what a valid transaction looks like and what a fraudulent transaction looks like.

On another front, Mastercard is increasingly using AI in conjunction with biometrics — such as fingerprint, iris and facial recognition — as another tool to verify the identity of card users. For example, the Mastercard Identity Check service allows online shoppers to authenticate a purchase by touching the screen of a smartphone or simply showing their faces to the device and blinking — a concept sometimes referred to as “selfie pay.”

And now Mastercard is taking biometrics to even higher level with the capabilities of NuData Security, a passive biometrics and behavioral analytics company that Mastercard acquired in 2017. The NuData technology identifies and verifies users based on their online interactions — behavior that can’t be replicated by a third party. For example, the technology considers how individual users hold a device, the way they swipe it and tap it, and the pressure they put on the screen.

“These are all unique signatures,” Curcuru says. “We use machine learning to build that profile. And it continuously builds that profile over time. Machine learning knows it’s you from your signature.”

In cases in which a user’s biometrics don’t match the profile developed by machine learning, artificial intelligence steps in and stops the transaction.

On the backend, lightning-fast identity verification is enabled by the close coupling of compute and storage, which reduces system latency. For this use case, Mastercard leverages Dell EMC PowerEdge servers and Isilon scale-out network-attached storage systems, which allow the close coupling of compute and storage while enabling performance and capacity to scale independently of each other.

Ultimately, when it comes to fraud prevention, the name of the game is to outsmart some very clever people with criminal intent. “Fraudsters are now automating their fraud,” Curcuru says. “They are brilliant.”

True, those criminal minds may be brilliant, but not nearly as brilliant as the AI systems and biometric tools in the Mastercard arsenal. Thanks to those systems and tools, billions of legitimate cardholders can use their payment cards with the confidence that Mastercard has the security issues covered.

For the full story, you can watch an on-demand replay of the Curcuru’s presentation from the “Unlock the Power of Data” virtual event. The Mastercard content begins around the 18-minute mark in the larger presentation. Read more in Fighting fraud the smart way – with data analytics and artificial intelligence.

Copyright © 2018 IDG Communications, Inc.