Economic crimes have been plaguing the Financial Industry with increasing cost each year. The PricewaterhouseCooper global economic crime survey for 2016 suggests that about 36% of organizations have experienced crime. Failing to properly identify and prevent fraud is an expensive proposition that costs the financial industry billions of dollars per year.
In 2011, fraud stemming from economic crimes, cost the financial industry approximately $80 billion and U.S. credit and debit card issuers alone lost $2.4 billion. A recent Nilson Report estimates that losses around the world, due to credit card fraud exclusively, would be more than $27.69 billion in 2017, which is about 11 percent more than the previous year. While the financial impact for institutions can be significant in monetary terms, the experience, for individual victims of fraud can be costly and even lead to identity theft, which can take years to resolve. Retail Industry is one of the most affected industries for fraud with Point-of-Sale systems being most vulnerable to fraud.
To combat this, Loss Prevention departments of many retail stores have installed Close-Circuit-Television for monitoring and recording fraudulent activities. The advent of increased technology adoption and Internet has fueled these crimes, with more opportunity for theft and misuse of credit card and Personal Credit Information (PCI) data. Advancements in software technologies and the growing use of e-commerce platforms have exponentially increased the risk of fraud for financial services companies and their customers. The internet has become the busiest ordering place in the world for everything under the sun – from food to electronics to hotel and travel reservations.
The opportunity to do business is immense for retailers that offer delivery of goods online, but the potential for fraud is several times higher than retailers in brick-and-mortar stores. According to a survey, Pymts.com estimates that online fraud accounted for 3.4 percent of all e-commerce sales in the 3rd quarter of 2015, and about 2.1 percent of all transactions.
Current fraud detection systems operate on a set of rules or algorithms, unique to the institution and the domain or activity. This could vary from simple rules, such as flagging cash withdrawals from ATMs that are more than a certain amount or purchases done using a credit card that take place outside state or country of the usual geo-location of the card holder. However, these fraud detection systems are not efficient, fast and are prone to false positives and false negatives. Over the past decade, the speed of financial transactions has been increasing at a tremendous pace with the world wide shift to immediate payment systems that mandate faster fraud detection systems. Fraudsters have increasingly and continually adapted and changed their tactics to find loopholes and avoid current fraud detection systems. This requires modern fraud detection systems to constantly improve their algorithms and update their techniques, which is rather difficult.
Artificial Intelligence, specifically Machine learning, is perfectly suited to addressing fraud in financial transactions, partly because of the feedback loop that is inherently present in such financial transactions. Machine Learning relies on data that can be used to train itself to achieve the required goal. When a fraud attempt is successful, is can be used to train the system using Machine Learning to detect such transactions in the future. Thus the system will automatically learn to improve it’s fraud detection capabilities, thereby making it difficult for fraudsters to keep up. Machine learning systems can analyze vast amounts of historical data to identify patterns associated with fraud, which are otherwise difficult to detect, taking into account many more data points than would be possible with manual methods alone, including detailed patterns of behavior associated with specific accounts.