It’s an eternal worry for banks: how best to detect and report money laundering. And the problem grows bigger and costlier by the day. Such transactions represent 2 to 5 percent of global GDP—or roughly $1 trillion to $2 trillion a year, according to a PriceWaterhouseCoopers global survey of banks in 2015 and 2016.
And to be sure, the burdensome endeavor to comply with anti-money laundering (AML) regulations finds banks fighting a battle on two fronts: to satisfy regulators and head off financial consequences (including big fines) before they occur. But financial services organizations now have a powerful new weapon at their disposal: artificial intelligence, or AI.
For some time, regulators have pressured institutions employ more sophisticated technology to help battle money laundering, according to Julie Conroy, research director at Aite Group, a research and consulting organization based in Boston. Banks for their own part are driven to embrace new technology by a desire to ease the unrelenting cost pressures of compliance.
While a long overdue development, good reasons exist for the tech delays. “AML has always been quite frankly behind their anti-fraud counterparts in analytics” she explains. This is due to the heavy emphasis regulators place on model risk management, especially in the U.S., Conroy points out.
But the tools are on their way. Since November 2016, IBM Corp. has applied the cognitive technology it developed with Watson to the task of complying with complex financial regulations. That’s when the company acquired the Promontory Financial Group a financial regulatory advisory company in Washington D.C. “So, this modern technology is just entering the marketplace right this minute,” says Gene Ludwig, founder and chief executive officer of Promontory and former Comptroller of the Currency.
Watson is already proving the benefits of applying cognitive learning to AML compliance, according to Ludwig. It has some key advantages. First, he says, the tool is highly precise and can detect patterns that expert human monitors might occasionally miss.
Also, it can tap into data from all over the world, allowing it to spot patterns that might be unclear from relying on data sources within one organization. Finally, it learns not just from what AML regulatory experts at Promontory teach it, but also from its own experience in spotting trends.
“It just gets better and better,” says Ludwig.
Because the cognitive learning tool has so much computing power, it is faster and its cost savings are commensurately larger, according to Ludwig. The imperative, meanwhile, has never been greater.
The Clearing House in New York issued a report in February that calls for a new and more effective regulatory and enforcement paradigm for anti-money laundering—and for countering the financing of terrorism. The current state of affairs represents “a significant misallocation of resources away from activities that would be the most productive,” says Greg Baer, president of The Clearing House.
The report also notes that the AML compliance burden leads banks to de-risk by shying from customers who might pose more of an AML or terrorist-funding risk. In the process, they push more accounts and activity out of the best-regulated banks, and off to institutions where they will no longer be properly monitored.
“Banks can lose a ton of money in fines and penalties if they are caught in situation where they allowed transactions to go through that had not be executed with the correct amount of due diligence,” says James McGovern, partner at Hogan Lovells in New York. McGovern is also a former chief of the Criminal Division of the U.S. Attorney’s Office, Eastern District of New York, and also deputy chief of the Business and Securities Fraud Section.
In 2015, for example, more than 28 percent of enforcement actions issued by banking agencies against financial institutions were for AML and Bank Secrecy Act compliance, according to an analysis by Sullivan and Cromwell in New York.
And that’s not the only area where banks feel a bottom-line effect. To deal with the ever-rising costs of complying with anti-money laundering (AML) regulations, banks by necessity must work diligently to improve the effectiveness of their crime-fighting efforts.
Finding new ways to contain costs has not been easy. Fighting money laundering crime is a big and growing part of the banking business—and those costs are expected to mount at a cumulative rate of 2 percent a year. Projected through 2020, the jump could hit 10 percent overall: from $4.1 billion in 2015 to $4.5 billion, according to GlobalData plc, a data and insights solution provider based in London.
It’s also a labor-intensive enterprise and requires sophisticated monitoring capabilities. When banks spot potential money laundering or fraud, they must file Suspicious Activity Reports and Currency Transaction Reports with Financial Crimes Enforcement Network (FinCEN), a bureau of the U.S. Treasury Department.
For a bank to perform its proper due diligence, it needs to “do deep dives” to get more information to build more support of any decision to accept the client. This extra effort can be very costly, Andrews says. Meanwhile, a form of strict liability exists for the bank, in terms of fines from regulators, for failing to screen properly should something go wrong later.
But AI can alleviate some of those problems—and save money—by reducing the instances of false positives. That in turn reduces time wasted on bogus alerts and the number of mistaken Suspicious Activity Reports that a bank is required to file.
It the end this increases the accuracy of filed reports, Ludwig says. And spending less time on false positives creates a true positive for banks: “It makes it easier to spot the bad guys.”
Compliance training and professional development courses that are efficient, effective and on-point. Give your people the latest industry-approved tools they need to improve performance, reduce operational risk and better serve your customers.