CUSO-News---Payments-Report

close

Categories

More Tags

Subscribe to Email Updates

Popular Stories

Pay-by-Bank: Anticipating the Next Wave of Innovations
Understanding Enumeration Attacks and How to Prevent Them
What It Means to Have a World Class NPS
Travel Disruption Just Became Easier with Visa's Digital Card Replacement
Combatting AI-Powered Fraud wtih AI-Powered Fraud Prevention
Written by Cyndie Martini
on August 27, 2019

Using AI for credit card fraud detection is nothing new. The majority of credit card providers now use AI for real-time fraud detection. Visa has gone a step further with their AI implementation. In this article, we'll look at what Visa's new AI fraud implementation is all about and why it is so different from the competition.

Most credit card AI fraud detection algorithms use Machine Learning. Machine Learning requires historical data to train an AI algorithm. Once deployed into the wild and using what it learned from historical data, the algorithm can flag similar suspect credit card transactions as potentially fraudulent.

How does Deep Learning differ from Machine Learning? Actually, Deep Learning is a subset of Machine Learning. Have you ever seen Russian dolls? These are the ones where you open the largest doll and there is another smaller one inside and on and on to the tiniest and last doll The outer most doll is AI, next is Machine Learning, and then Deep Learning. 

Al credit card fraud uses Machine Learning but only a small number of credit card processors use Deep Learning, like Visa. Machine Learning requires some setup from its programmers in order to function. Data has to be identified and categorized; algorithms have to be structured so that problems are solved in steps. Deep Learning doesn't require those two steps. Deep Learning also has a secret advantage called a neural network, which requires massive amounts of data with powerful computers in order to execute efficiently. With only a small amount of data, Deep Learning will not perform as well as traditional (Machine Learning) algorithms. Another advantage of Deep Learning is that it is quicker to train and executes much faster than Machine Learning algorithms, which are critical when trying to identify credit card fraud in real-time.

For Visa, Deep Learning allows them to identify fewer false positives and work with more complex patterns. “It’s a massive breakthrough for us,”  Rajat Taneja, executive vice president of technology and operations for Visa, told The Wall Street Journal.

What does this mean for smaller banks and credit unions who use white label Visa credit cards? It means more security for your customers. Card card merchants receive all of the benefits from Visa's heavy lifting.

Let Us Know What You Thought about this Post.

Put your Comment Below.

You may also like: