Fraud has become the most common crime in the digital world, with criminals making billions of dollars and banks paying billions in fines. As money launderers have increased their level of sophistication, their transactions have become harder to detect – and banks are having a harder time responding to their threat.
Understandably, banks are increasingly investing in anti-money laundering (AML) solutions, with artificial intelligence (AI) and machine learning (ML)-based solutions in particular gaining significant traction. The positive side of this is that AI and ML technologies have a lot to offer financial institutions in the AML fight. Unfortunately, many of the solutions that claim to be AI-driven don’t use true AI – vendors just tack on minimal AI for marketing purposes. To combat the increasing scourge of money laundering, that’s not going to fly. Banks must have an AML solution that demonstrates AI’s true capabilities.
The necessity for action has increased with the passage in the U.S. of the Anti-Money Laundering Act of 2020. Many consider this to be the most substantive AML legislation that Congress has passed in decades. The new law contains higher fines for violations, greater shielding for whistleblowers and more authority for regulators in America to obtain documents from financial institutions in other countries.
Changing times, changing fines
The pandemic and its lockdowns introduced new fraud pathways for financial criminals and money launderers. Because financial institutions had to do more business online, the suddenly remote workforce enabled yet more possibilities for exploitation of financial and enterprise cybersecurity vulnerabilities. The converging realities of increased online and remote banking and federal stimulus assistance have caused a significant uptick in financial fraud and associated money laundering.
As money laundering has changed, so has enforcement. In 2020, U.S. banks had to pay $14.2 billion in fines; most of that was due to AML violations. For comparison, they paid $8 billion total in 2019. The passage of the new federal money laundering rules begins a new season of AML regulation and enforcement for America’s financial institutions. The possibility of stronger sanctions and higher financial penalties provides renewed motivation for financial services organizations to comply with AML best practices.
Old-school AML solutions can overwhelm
Even though banks have increased their AML investment, the problem remains. AML compliance in U.S. and Canadian financial institutions cost $31.5 billion in 2019. The study discovered that a layered approach to compliance yields better results, particularly if they contain AI/ML. However, many institutions continue to depend on a technology-manual hybrid for AML compliance, and that’s not ideal for cost or performance. Manual effort leads to an overwhelming number of false positive alerts, which requires financial firms to keep numerous investigators on staff. The majority of their job involves double-checking completely normal financial behavior. Thankfully, true AI puts an end to this expensive and inefficient process.
What sophisticated AI looks like
Supervised machine learning is a technique that many of today’s AI/ML solutions depend on. Basically, the AI receives information about things that are true and then looks for new examples. The ability to recognize dogs is an example. The AI sees multiple images of dogs, and then it sees images of things other than dogs; this enables the AI to “tell” whether an image is “dog” or “not dog.” However, in terms of telling the difference between what is “crime” and “not crime,” identification becomes trickier.
Instead, financial institutions need a type of AI that’s more advanced and has the ability to map behaviors within a system. Imagine working at a bank serving 8 million customers, and each of them does thousands of transactions each year. One customer sends money to another customer, who then sends money to a third customer. This happens repeatedly throughout the year, with varying amounts and varying kinds of transactions. In the midst of all these transactions is money laundering, but who could find it? There’s no way for an investigator to know what to look for ahead of time.
Banks can begin to discover these hidden crimes if they use the right mix of techniques. An important first step is to find and map the many and varied behaviors of different customers. Today, ML technology exists that is able to learn about financial behavior and the flow of money, then look at all the different customers’ behaviors compared to each other. This will naturally unearth higher-risk behavior. Banks have a historic, investigation-based view of risk that they can overlay on the ML-generated map so the system can zero in on specific areas and create alerts. An automated solution like this supports regulatory review and achieves high risk coverage and accuracy.
Lower fraud, lower risk
One multi-billion-dollar industry, money laundering, has given rise to another: anti-money laundering. Criminals’ schemes have become so sophisticated that vendors have incorporated AI and ML into their solutions. However, standard ML techniques are not up to the task at hand. AI is not all the same; solutions that automatically map behaviors are the ones that will help professionals at financial institutions spot crime. It will also reduce fines and risk, as well as human effort. Banks now have a tremendous opportunity to expose money laundering as well as improve their risk profile by choosing sound AI technologies.
Written by Dr. Stephen Moody.
Track Latest News Live on CEOWORLD magazine and get news updates from the United States and around the world. The views expressed are those of the author and are not necessarily those of the CEOWORLD magazine. Follow CEOWORLD magazine on Twitter and Facebook. For media queries, please contact: firstname.lastname@example.org