credit cards By Chase Rimmer / September 11, 2018 Credit card fraud losses in the US amounted to $3 billion in 2014, a figure that is projected to decrease to $1.8 billion in 2018, says Statista. The development of better techniques to detect fraud is expected to reduce the number of anomalous transactions. It also aims to help banks and financial institutions deliver a better customer experience to credit card holders. Specifically, the utilization of artificial intelligence and big data are anticipated to detect fraud using algorithms fed to machines, track transactions in real time and reject dealings if they fail to pass the probability number indicator. Big Data and Its Application to Credit Card Fraud The concept of gathering big data is nothing new, having been in existence since the 1940s when it was first developed by Hari Seldon. However, the large volumes of harvested information have found new applications and uses. Data is used for many purposes, whether to provide input for decision-making or to predict behavior. Whereas before, flagging credit card transaction fraud would have involved traditional data analysis techniques and further validation performed by humans; today, machines and artificial intelligence do the job. To prevent, limit or slow down credit card fraud, an algorithm detects suspicious activities that would send alerts to the companies that issued them. Fraud detection covers all types of credit cards, regardless of the differences in perks offered by each kind. Protection against fraudulent transactions involves the application of artificial intelligence and big data resulting to efficiencies. Credit Card Ownership and Fraudulent Activities Without a doubt, credit card ownership in the US has grown tremendously over the years with over 174 million Americans in possession of at least one card in 2015 according to the Census Bureau. The volume of transactions has also increased significantly making it harder to keep track of discrepancies. In fact, credit card losses worldwide in 2016 reached a whopping $24.71 billion as estimated by the Nilson Report. An identity theft is committed every 2 seconds including credit card fraud according to the Javelin Strategy. This is where AI and big data step in. By training machines to learn an algorithm which consists of many factors, they will become familiar with typical transactions of cardholders. Battling the Bad Guys Transaction sequencing will help machines determine normal buying behavior, such as typical purchases in a grocery store, gas station or fast-food outlet. The algorithm is adjusted to account for factors such as the reputation of vendors, IP addresses, locations, card owner’s purchasing patterns and so on. Aberrant activities are detected when real time transactions fail to meet the probability number cut-off that is assigned when all factors are considered. In this case, the transaction is aborted or rejected. Human interference is minimal, although the company can still call the cardholder to verify if they have made the erroneous transactions before taking serious action such as canceling the credit card. Fraud detection is like a race to outwit thieves. At present, advances in technology are helping prevent thefts. AI and big data are changing the way businesses improve customer experiencing by reducing credit card fraud transactions. In the process, companies avoid significant losses that will help improve their margins. Franck V.