Machine learning algorithms clearly afford a lot of opportunity to improve cybersecurity. CA Technologies is now putting that theory to the test in the form of a CA Risk Analytics Network that makes use of machine learning algorithms to identify fraud involving the use of not present credit card transactions.
Credit card companies lose billions of dollars a year when criminals employ stolen credit card numbers to buy any number of goods and services online. Terrence Clark, general manager for CA Technologies Payment Security solutions, says CA Technologies can now reduce those losses by as much as 25 percent. Given the amount of fraud being perpetrated, CA Technologies estimates CA Risk Analytics Network could save credit card companies over two billion dollars.
Delivered as a service managed by CA Technologies, Clark says the goal is to increase fraud detection in a way that does not impede legitimate card transactions. Clark says CA Technologies can achieve that goal by applying machine learning algorithms, neural networks and advanced analytics to identify which devices are being employed to conduct those transactions in less than five milliseconds. That approach winds up not only being more efficient than tracking individual credit card numbers, it provides a better customer experience, says Clark.
“There has to be less friction for the customer,” says Clark.https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=i
Clark says new specifications such as 3-D Secure that address authentication and security for card-not-present transactions using smartphones, mobile apps, digital wallets and other forms of digital payment will also go a long way to reducing credit card fraud.
Naturally, it’s hard to say with absolute certainty that credit card fraud can be eliminated. But as the ability to apply machine and deep learning algorithms against massive amounts of data in real time continues to improve, the opportunity to commit fraudulent credit card transactions should be sharply reduced.