We live in a world where terrorists and terror attacks have become a part of our reality more than we would have liked. Every now and then there are headlines that make us cringe to our very stomachs and make us wonder if there is a way out of it.
One such organisation that is making things worse for humanity is Daesh. Previously being run on an international level, the organisation is no longer able to carry out large-scale attacks, the group has now resorted to employing individuals, or small groups, to carry out the terroristic misdeeds for them. The employ of small cells and lone operatives have made it a difficult task for law enforcement officers to identify and stop such attacks in advance. This means, tracking of their funds is becoming tricker day by day.
Financial institutions have been employing anti-money laundering systems to catch suspicious activities for quite sometime, but since 9/11, those tools have also been employed to net terror-related transactions. However, with small amounts of flowing in from the several group to individuals from all around the globe, banks have been having a tough time following the trail.
This is where machine learning is coming to banks’ help. The tech is helping them in mining their internal data and find anomalies, which would have otherwise gone unnoticed.
The West has numerous regulations and legislations in place since the 60s where financial institutions are obligated to assist government agencies in detecting money laundering cases. However, even if the software being used is automated, there’s a possibility of even the best of the tech churning out false positives. According to a survey done by Dow Jones of over 800 anti-money laundering professionals, around 50 percent aren’t confident in the efficacy of automated screening processes.
The main challenge that the banks are facing is that are infinite possible permutations of transactions out there to consider in a rules-based system.
If a person from Belgium wishes to withdraw money from an ATM in Brussels, they would receive a Western Union transfer in Tunisia, and will have to use a credit card in Syria. They could also have an option of taking a payday loan or transfer money to their family. While individually, these activities don not raise any false alarm, but when carried out collectively, they create a pattern which artificial intelligence can gauge if there’s something questionable.
In addition to this, when AI is fed information such as financial histories of known, past terrorists and criminals from the world’s biggest banks, they can quickly find out any financial anomalies like how quickly the money is being moved around, where it moved, and how much was transferred, that might be occurring. Further, AI can also look into number sequences anomalies. When criminal minds are laundering money, they often end up falsify invoicing to make it look as a legitimate transaction. However, sometimes they end up forgetting the identification numbers of these invoices and end up making mistakes. Such mistakes can be easily identified through the help of AI and machine learning.
Such methods were employed to unravel the alleged drug trafficking ring in Panama — Grupo Wisa, a holding company which runs duty free stores in Latin American airports. QuantaVerse noticed that a series of invoices for large, round dollar amounts were being passed back and forth between businesses that had the same owner on the deed. The development helped in the arrest of Nidal Waked, one of Grupo Wisa’s proprietors for money laundering charges.