Banks process an incredible volume of transactions every day. How can banks identify white collar crime in an expected, streamlined manner without impeding legitimate transactions? Artificial intelligence and machine learning are the answer both today and in the future, but applying them may be easier said than done.
In the United States alone more than $14 trillion are routed through banking systems on a daily basis. From checking and savings accounts to mortgages and microfinance loans, there are a plethora of mediums that fuel the global flow of cash. This begs the question – how do banks detect illegal and fraudulent transactions amidst real ones?
While banks are working tirelessly to outpace white collar criminals, they run the risk of not seeing the forest for the trees. Hyper-specific detection rules may stop one route of money laundering, for example, but unless you have a system that can expediently read, react, predict and prevent fraudulent transactions, criminals will be able to adapt and find new ways to circumvent bank monitoring. Enter machine learning and artificial intelligence tools, which have the potential to reshape the compliance landscape using graph analytics and predictive modeling in both front and middle-office banking operations.
Many financial institutions are still undecided as to whether these innovations are worth the risk of overhauling their current systems. However, the technology makes too much sense to ignore any longer; it could save banks millions of dollars and clean up errant funds allocation at a rapid pace.
Don’t fear the rise of machines
The case for integration of these innovations can be easily made after taking a close look at the current landscape. Systems for financial crime identification and reporting are both antiquated and inefficient.
For example, the process of pinpointing money laundering attempts should be straightforward, but it tends to be an exhaustive, time-consuming exercise. Today, banks utilise rules-based, IF-THEN systems to identify threats – i.e. ‘IF a client executes several $10,000+ transactions THEN they are flagged as a money laundering risk.’ If somebody breaches a rule, then a non-compliant flag is raised and the customer is investigated further to determine intent.
One clearly inherent problem arises with this rules-based approach – banks end up casting such a wide net that they wind up catching far more good guys than bad. Magnifying this is the fact that an entire corporate division, often times exceeding 1,000 employees, must dedicate their time to separating the wheat from the chaff. This is without mentioning that by the time a rule is perfected, the most malicious of offenders have already found a way to circumvent the identification process.
It’s no coincidence that roughly 80 per cent of compliance costs are spent on people, not technology. This is exactly why AI and machine learning adoption will result in significant savings on revenue and time for key banking departments. These systems produce models that can quickly comprehend user behavior while utilising predictive analytics to outpace defrauders at a remarkable pace.
The barriers to AI and machine learning implementation
Machine learning has the potential to remedy a currently inefficient system, but it’s not without its hurdles. The key component here is that there is tremendous pressure on banks to work with regulators in identifying an amicable solution for both parties that catches more defrauders and fewer bystanders.
From the POV of banks, these new systems make sense—less money spent on evaluating threats and faster execution—but they pose a significant regulatory challenge. When a money laundering attempt is reported, the regulatory authority needs to be made privy to every pertinent detail of the activity. Utilising a rule-based system allows for full visibility of the audit trail, but machine learning tends to come as a black box solution. In other words, you can see the input and output of a bank’s vetting process, but you can’t see how they got to the final resolution.
It is imperative that we transition from traditionally black box machine learning solutions to a white box approach. We can’t feasibly expect to rapidly implement cost-effective machine learning initiatives without taking into consideration the inherently granular needs of regulators. As the pace of innovation increases, we’re closer to a harmonious solution every day.
So what’s next?
There is no one size fits all solution here – machine learning tools need to be thoroughly tested to ensure we’re reaching an agreeable outcome for both financial institutions and regulatory agencies. For example, companies are already working toward machine learning approaches that are fully explainable while working more closely with investigators on what information they need is always invaluable. The onus is on both regulators and banks to ensure that these technologies and their intrinsic risk are managed.
While the road to widespread adoption of machine learning systems is an arduous one, the benefits extend beyond identification of potential malicious activity. Whenever a bank identifies suspicious activity, they need to develop an extensive, hyper-specific report for regulators to review. What if AI auto-populated these documents, saving an inordinate volume of man-hours each day? Add this to the time already saved from a non-rule-based architecture and we’re looking at a significantly different financial landscape.
It’s absolutely a certainty that very soon, machine learning in Regulatory Compliance will be a necessity rather than just another option. Machine learning and other branches of AI present an unprecedented opportunity for banks and regulators to come together in reducing regulatory burden but at the same time proactively preventing more illegal activity.
Ambreesh Khanna, group vice president and general manager, Oracle Financial Services' Analytical Applications
Image Credit: PHOTOCREO Michal Bednarek / Shutterstock