Despite ongoing efforts to curb money laundering schemes, many organizations still have a difficult time keeping pace with the sheer volume of transactions taking place.
Traditional rules-based anti-money-laundering (AML) solutions only add fuel to the fire, generating countless false positive alerts that leave organizations overwhelmed and dealing with costly mistakes.
Artificial intelligence, according to information technology and services firm CSI, will help organizations not only deal with money laundering frustrations but also catch any potential red flags before the business is affected. In its recent “Anti-Money Laundering (AML) Growing Pains” white paper, CSI outlines just how much AI-powered AML software can help businesses adapt to evolving money laundering strategies while also reducing operational costs. By analyzing historical AML data, both internal and external, the technology can identify anomalous activities and connections that rules-based systems often miss.
The Current State of Money Laundering
The relentless flow of money laundering poses a significant threat to financial institutions. According to the United Nations Office on Drugs and Crime (UNODC), roughly 2% to 5% of global GDP—approximately $5 trillion in 2022—is money laundered.
Financial institutions struggle to keep up with persistent money launderers, who are always one step ahead, for several reasons. For one, there’s a lack of resources available to help organizations build better lines of defense. Budget constraints also limit many. As a result, organizations fail to implement effective internal controls to monitor and report suspicious activities, resulting in costly fines and regulatory penalties.
Any businesses involved with moving money need to pay attention to AML laws. If they don’t, they’re at risk of facing fines. According to FinCEN statements analyzed by CSI, many organizations, including two large depository institutions, a community bank, and a perfumery faced recent fines. In one case, the white paper noted, “FinCEN imposed a $100 million CMP for what it described as a ‘willful’ failure to implement a program meeting all the recruitments of AML compliance.”
The Problem With False Positives
Until recently, rules-based AML solutions were the most sophisticated tools available. However, they allow an organization to implement only up to 10 rules—and given that the rules are standardized, money launderers have figured out loopholes.
Rules-based AML solutions can be a double-edged sword because although they aim to detect money laundering, they also generate a high volume of false positive alerts. These alerts require manual analyses, which are time-consuming and prone to human error.
A large depository institution that has leveraged rules-based AML solutions (averaging 4,500 daily alerts) told CSI that it has had difficulty vetting all the alerts. The institution employed 10 AML analysts who work eight-hour days, and each team member “needed to either clear or escalate 56 alerts per hour.” That leaves each team member with approximately one minute to investigate every alert that comes through. Understandably, this has left the workers behind in their work, unable to keep up with the demand.
How AI-Powered Solutions Can Help
AI-powered AML software can help, particularly when leveraging machine learning models to adapt to ever-evolving money laundering tactics.
“An AI-powered AML solution can more easily spot layering activity meant to hide money laundering,” CSI noted in its white paper. “With rules only, a sudden burst in account activity creates an alert, but it is difficult for an AML analyst to determine whether it should be cleared or escalated. Not so, when their dashboard visually shows the connection between counterparties, such as similar amounts and usage texts, topped by passthrough activity.”
AI can also close alerts that can be ruled out based on learned patterns, which reduces false positives and enables AML investigators to focus on high-risk cases. What’s more, the technology analyzes extensive data to create risk profiles and scores for accountholders, making it easier for AML analysts to prioritize alerts and investigations.
Because there are a lot of intricacies to learning about specific patterns—including geographic locations, politically exposed persons (PEPs), or the length of someone’s account—the more time AI spends analyzing the data, the quicker the technology is able to catch potential money-laundering schemes. Overall, CSI points out, AI is able to learn and adjust for potential risks a lot faster than an AML analyst can.
Automatic case-closing functionality may be one of the key perks of AI-powered AML solutions. It conducts the first round of reviews, automatically closes cases that the model doesn’t think are fraudulent and passes on the remaining cases to the analysts. This cuts down on workflow substantially. Organizations that have leveraged auto-close functionality generally see 70% fewer false positives, per CSI, noting one particular use case where a payments technology processing company saw a 95% reduction in false positives, which saved the company 20 full-time employees.
Conclusion
Traditional rules-based AML solutions, while well-intentioned, have struggled to keep pace with money launderers, who are always one step ahead.
AI-powered AML software systems can be a game-changer in the ongoing battle against illicit financial activities. By leveraging machine learning models to continuously adapt and learn from historical AML data, organizations can better identify suspicious activities and connections that rules-based systems often miss.
Financial institutions that have embraced AI not only safeguard themselves against the persistent threat of money laundering but also benefit from streamlined operations and reduced regulatory risks. As the scale of money laundering grows and diversifies, AI proves to be a valuable ally in the fight against financial crime.