Hear from Michael Thelander, fraud and authentication expert and author of Multi-factor Authentication for Dummies, about machine learning in the context of fraud and authentication for the financial services industry.

“We understand what machine learning is, so what are some of the ways that we can apply machine learning, particularly in the finance sector?

There are two areas in particular: Fraud prevention and detection in one, and then authentication in the other one. With machine learning determining what is normal, giving us an idea of what normal activity and behavior look like, it becomes very easy for those abnormal, non-normal activities and routines, those high-risk activities, to pop out without having to create rules for them in advance. Another way, in authentication, is when we start recognizing our customers’ returning devices, we can use machine learning to determine what is the normal or expected rate of change in that device over time. If we haven’t seen that customer’s device in the last two weeks and we see it now presented through our website, how much change should we expect to see in that device? And if it’s appropriate to the level of change that we observe in that device, it may be that we just let them in with a low friction authentication experience. If there’s a great deal of change, based on what machine learning tells us the average or the normal change is, we might then drive it through a step-up authentication process. So machine learning gives us intelligence to provide a better user experience in that case.”

For more information, check out our blog. Or read the Aite report on how machine learning is becoming a competitive issue in finance.