THE BIG IDEA: PayPal uses Deep Learning to Detect Fraud

undefinedSo we all know PayPal, but just in case you have never heard of the service or ever used it in the past, allow me to let you in on what they do. PayPal is a service that allows you to pay for goods and services, or send money to another person or party, all digitally. If you have a confirmed email address, you can set up a PayPal account and send money anywhere globally. It’s secure, safe, and all done online. You have the ability to also carry out transactions over your phone as well with the PayPal app. It’s been around since—and this might surprise you—1998. It’s a very convenient service.

Now as October is National Cybersecurity Awareness Month, let’s talk a little about identity theft, fraud, and how PayPal works to remain secure. I don’t doubt with recent hacks on Ashley Madison, Apple, and just reported today T-Mobile, you may be a little leery of entrusting your bank details to a site that describes itself as “secure” when there is, obviously, no real way to assure clientele that; and banking information is gold for identity thieves looking to make a quick profit. Truth be told, there is no way to guarantee complete and total security online, but there are companies that are doing everything they can—and getting pretty innovative at protecting, you and your data.

Artificial IntelligencePayPal is currently using Deep Learning to detect fraud. What Deep Learning is exactly could be a Big Idea all on its own, but in a nutshell it is a network of computers all working to “learn” behaviors and patterns so that it will recognize those patterns and deduce a conclusion. It’s a bit like when you are on Facebook and follow a link to Amazon or Etsy, and then on returning to Facebook your right-hand sidebar is suddenly filled with of ads from Amazon and Etsy. That’s Deep Learning, and PayPal is using it to keep you clear of fraud. See, one of the problems with fraudsters is that they will try different things to get to you private data or, as it is known in cybersecurity, Personally Identifiable Information (PII). You can, of course, write an algorithm—a series of commands and procedures that computers recognize—to detect what a fraudster might do. However, if the algorithm fails or the fraudsters change whatever they are doing, you need to use something else, so you have to create another algorithm and that takes time, time that hackers utilize to access a network. What you need is something that adapts to what fraudsters do or will do. This is what Paypal is implementing: Deep Learning (DL). In a sense, PayPal has created a network of computers patterned after neural networks in our own brains. Just like our own neural networks, PayPal’s DL network is constantly looking for patterns or correlations to user’s behaviors. By using DL PayPal can see if there is something unusual, flag it as a concern, and then either have one of their own experts or the account owner themselves take a look and see if this is approved behavior or not.

Paypal has been doing this DL approach for quite some time and have been successful enough with it that many other companies are considering artificial neural networks as a way to detect cyber-crime. One company, Palantir Technologies, used a DL network to locate terrorists, a very exciting method in keeping more than just your PII safe.



shurtz.jpgA research physicist who has become an entrepreneur and educational leader, and an expert on competency-based education, critical thinking in the classroom, curriculum development, and education management, Dr. Richard Shurtz is the president and chief executive officer of Stratfdord University. He has published over 30 technical publications, holds 15 patents, and is host of the weekly radio show, Tech Talk. A noted expert on competency-based education, Dr. Shurtz has conducted numerous workshops and seminars for educators in Jamaica, Egypt, India, and China, and has established academic partnerships in China, India, Sri Lanka, Kurdistan, Malaysia, and Canada.


Leave a Reply

Your email address will not be published. Required fields are marked *