Why machine learning is vital to moving one step ahead of fraudsters

Image Credit: Shutterstock (Image credit: Image Credit: Gustavo Frazao / Shutterstock)

How do you consistently hit a moving target? That’s the challenge facing fraud prevention teams as criminals set the sights on credit cards, current accounts, loans and other financial products.

Experian’s Hunter fraud statistics show that younger people – particularly millennials, many of whom live in blocks of apartments – continue to be the most common victims for fraudsters, yet criminals are also looking elsewhere.

Generally, we are seeing fraud against older people rising. Rural dwelling, wealthier homeowners, who make lengthy commutes to work or are retired, experienced nearly a 30% increase in fraud last year. Meanwhile, ‘empty nest’ or families with older children in suburban areas saw a 7.5% rise in fraud against them 

At Experian we identify a new fraud every 15 seconds. However, to each individual victim of fraud who experiences the upset and inconvenience of a criminal abusing their identity, it’s little comfort. Equally, organisations are keen to stop losing money to fraudsters, while causing minimal inconvenience to the customers they serve.

Fighting fraud with machine learning

Machine learning has become an essential tool in the fight against fraud.  

Fraud prevention systems have typically been rules-based, understanding the risk a particular application poses before flagging those which arouse suspicion to a fraud prevention team for further analysis. Whether an application is marked up depends on the rules set and the organisation’s appetite for risk and disrupting a customer journey.

Machine learning technology goes further. It looks at the results of an application, whether it was found to be fraudulent or not, and then uses this information to make better decisions in the future. The models for machine learning are data hungry – the more information it has at its disposal, the higher quality decisions can be made.  

The unique advantage Experian has in the machine learning market is having the ability to access these known fraud datasets to work from, unlike other organisations which are starting from scratch and waiting for the data machine learning requires to improve end results. Further, at Experian, we have the advantage of using the expertise of some of the leading data scientists globally in our five regional DataLabs to find new ways to stay one step ahead of the fraudsters.

Putting machine learning technology in place will help organisations to become more efficient in their search for fraudulent applications. There will be fewer ‘false positives’, where a genuine customer is referred for closer inspection, allowing fraud teams to focus on a more concentrated batch of applications.  

More consumers will get to enjoy a frictionless experience when they apply for services, while organisations can have more confidence they are dealing with genuine customers. A good result for all parties – except the fraudsters.

Micah Willbrand, Product Director for Identity and Fraud at Experian UK&I

Micah Willbrand

Micah Willbrand is the managing director for Identity & Fraud at Experian UK and Ireland. He has extensive experience within global technology companies, working in commercial, consultative, operational and strategic areas. He has worked alongside gaming organizations for more than 15 years around identity, age verification, and fraud strategies.

His strength is unlocking value and growth in businesses through a unique ability to organize direct and indirect stakeholders on a vision whilst enabling structured and comprehensive management of day to day operational budget commitments.