Anti-Money Laundering (AML) in the Artificial Intelligence (AI) Era

By Susan Hu

Introduction

According to the United Nations Office on Drugs and Crime (UNODC), the estimated amount of money laundered globally in one year is 2 - 5% of global GDP, or $800 billion - $2 trillion U.S. dollars. As stated by Thomson Reuters, in the U.S. alone, financial institutions are currently spending about$8 billion U.S. dollars a year on AML compliance.

Financial institutions rely on AML systems to detect crime. However, as money laundering criminals are becoming more sophisticated with new technologies, banks are facing challenges to keep these systems effective. That's where Artificial Intelligence (AI) comes in to help.

Challenges the financial industry is facing

A recent Dow Jones-sponsored ACAMS survey reveals that false positives are one of the most challenging aspects of AML for bank compliance teams. Being able to reduce false positives would give financial institutions time and empower their human resources to focus on catching actual money-laundering criminals.

Historically speaking, to detect money laundering, financial institutions have always relied on rule-based models that flag customer activity. It's often based on whether a transaction triggers certain pre-set threshold or rules - often relating to dollar amounts.

As alerts are triggered, banks' AML teams manually analyze each alert to determine whether the activity is an actual threat and a suspicious activity report (SAR) with a regulatory body is needed. Most generated alerts are legal behavior that rule-based AML systems falsely determine. On the other hand, these legacy AML systems also overlook a significant volume of actual suspicious activities, often when they cannot be identified based on the established rules - the unknown-unknowns.

As money-laundering criminals are using new technologies to conceal their illegal transactions, banks' outdated systems are costly and inefficient. Banking and financial institutions around the world are facing increasing risks of over working its compliance team while leaving money-laundering transactions undetected and paying huge fines to regulatory agencies.

Graphen's AI AML

Graphen's AML Solution is built upon Ardi, an AI platform that incorporates advanced machine learning and graph computing to analyze relationships between different entities.

Using next-gen AI, Graphen's AML can detect relationships between account holders, transactions, companies, etc. These analyses are used to assign each bank customer with an aggregate risk score reflecting the risk that they are engaging in money laundering activity. The relative value of these scores enables compliance team to efficiently organize their workflow and focus on reviewing the most pertinent threats first.

Interface of Graphen AML Solution

The Graphen AML Solution integrates information from the institution, including Know Your Customer (KYC), Customer Due Diligence (CDD) information collected on customers and transaction records, as well as alerts generated in the bank's current AML system. It creates an aggregated risk score for each account. The system also monitors live transactions and name-screening, identifies transacting parties that are already on a watchlist or has relationships with high-risk entities. These functions enable Graphen's AML Solution to reduce the number of false positive alerts and find previously undetected unknown-unknowns, allowing a much more efficient AML detection and reducing the bank's risk of exposure.

Using Graphen's AML Solution, banks and financial institutions can still create money-laundering detection rules that will monitor customer activity and generate live alerts of rule-violations. Users can customize each rule by selecting parameters and setting values, or they can directly create and manage rules using scripts. This functionality ensures banks can work on the foundation of their legacy AML systems to generate necessary rules for maintaining compliance with relevant laws and regulations in their jurisdiction.

Conclusion

As money-laundering activities are becoming more sophisticated and deceptive, banks and financial institutions need to upgrade their current AML systems to fight financial crime in a much more efficient and effective way. Using advanced AI to compliment and update legacy systems is the future of AML.

If you want to learn more about AI, Graphen's AI AML solution, and how it can help you increase your AML accuracy, please reach out to me at susanhu@graphen.ai.