Credit unions exploring artificial intelligence to fight fraud

CO-OP Financial Services is using machine learning technology to help its members’ systems learn and adapt to new threats

A credit union network in the USA is using artificial intelligence to help prevent fraud.

CO-OP Financial Services, which supports its 3,500 credit union members with ATM, card payment and mobile services, has introduced a machine learning-based risk management tool aimed at fighting fraud.

President and chief executive Todd Clark said: “A key pain point for credit unions in 2017 is that fraud mitigation tools have not kept pace with the relentless advance of fraud. CO-OP Financial Services has an equally relentless commitment to credit union and cardholder security.

“We are investing US$20m this year on new innovations to ensure credit unions have access to the very best technology in the payments industry. Implementing machine learning capabilities is a key initiative in delivery on that promise.”

CO-OP is working on the project with Feedzai, a company specialising in artificial intelligence for payments applications. The two are aiming to co-develop a series of fraud-fighting machine learning advances, including the risk management tool.

Related: Could Amazon Mechanical Turk workers own AI technology?

The machine learning platform, which will be used by CO-OP member credit unions only, works in real-time, incorporates advanced analytics and provides alert and case management capabilities for fraud and risk services. Machine learning allows robots to automatically learn from data and adapt, with or without human supervision.

The platform will feed on historical and current data and use a strengthened transaction scoring model to better predict debit and credit card fraud, reducing credit unions’ losses while keeping false alarms to a minimum.

The tool will take in data to define transaction risk scoring, incorporating events such as cardholder PIN changes and purchasing patterns to better predict expected behaviours. The scoring will be updated in real time based on constant learning, which means that suspected transactions detected by machine learning can be declined or send to the fraud team members for analysis.

“Artificial intelligence has previously been thought of as technology only available to governmental organisations or the largest of corporations,” said Nuno Sebastiao, chief executive of Feedzai.

“Feedzai is pioneering ‘attainable AI’ – AI that will really work, right now for any sized company. There is perfect synergy between ourselves and CO-OP, and we are very excited to be working with them to bring machine learning to the aid of credit unions.”