The Shadows of Wealth: Applying AI to the Hidden World of Money Laundering
Applying artificial intelligence (AI) to counter global money laundering efforts poses challenges. The abundance of data can be overwhelming, and money movements are massive, nonstop and produce information trails by design.
Therefore, when implementing a system to detect wrongdoing, having the right scope is crucial.
“AI [artificial intelligence] can drive significant efficiencies in typical operational hotspots, such as customer due diligence, screening and transaction monitoring controls,” said Patrick Craig, financial crime lead at EY, an accounting and consulting firm.“Money laundering activity around the world is estimated to be between 2% and 5% of the global GDP.
This would amount to $800 billion to $2 trillion per annum.
Money laundering is a process where those trying to transact funds capitalize on opportunities to turn ill-obtained profits into legal funds that enter the banking system.
It is “the conversion or transfer of property, knowing that such property is derived from any offense(s), for the purpose of concealing or disguising the illicit origin of the property or of assisting any person who is involved in such offense(s) to evade the legal consequences of his actions,” according to the United Nations Vienna 1988 Convention.
“Money laundering follows phases: first, the illegal transaction that needs to be hidden; later, there is a layering process with different methods, and afterwards the money is concentrated again, and the procedure comes to an end,” said Mirlis Reyes Salarichs, professor at the Inter-American Defense College.
While the sum is large, these transactions are buried among troves of legitimate exchanges between lawful individuals. Employing humans to supervise this data stream is impossible.
“Hiding is precisely the most innovative and creative part of the laundering process,” Reyes Salarichs added.
Combatting illegal economic transactions is a security concern for most states, as criminal organizations, terrorist groups, adversarial states and other nefarious actors depend on illegal funding to sustain their activities.
To tackle the challenges of deploying AI solutions, Craig recommended beginning with a basic statement of objectives.
“The basis for developing AI-enabled [anti-money laundering] solutions must start with a clear statement of objectives to ensure that the design and implementation are aligned with the intended use and integrate effectively into business processes,” Craig said.
And the technology, by itself, yields few results if not accompanied by context.
“Without enhancing investigative and intelligence capabilities, AI will not yield significant benefits beyond incumbent controls,” Craig added.
Therefore, given that this area of law enforcement is riddled with subjectivity, for example, when setting thresholds for deeper inspection of an individual transaction, clear goals from the outset are crucial.
“Establishing clear performance indicators and parameters, which link to a well-defined risk appetite statement, will be critical to tracking whether the outputs from the AI are meeting objectives at an acceptable level of risk,” Craig said.
Another relevant issue is whether these technologies will be implemented in a business or a governmental organization. As with regulations around the world, enforcement has been privatized.
Protocols like KYC (know your customer) place liability on banks and other financial institutions for accepting illegal funds.
“There are no specific regulations about using AI for KYC, but authorities generally support firms experimenting with AI to strengthen compliance effectiveness,” said Moody’s Analytics, an investment intelligence company.
While liability falls on businesses, technologies seem to remain in their infancy.
“More than 95% of system-generated alerts turn out to be ‘false positives’ in the first phase of review, with approximately 98% never culminating in a suspicious activity report (SAR). High rates of false positives require manual reviews, which costs the industry billions of dollars in wasted investigation time each year and distract institutions from true suspicious activity,” according to a report by Google.
The company argued that setting rules to the system creates SARs that are irrelevant and informs wrongdoers of the parameters used for enforcement.
The company advocates a different approach, scoring transactions and taking an adaptive outlook that will be harder to trace for those under investigation.
Google declined to comment on these services.
Contributing reporter: Nuray Taylor