The scale of the problem has been demonstrated by the legislative activity of the European Parliament, which over the past five years has adopted three directives regarding anti-money laundering and terrorism financing.
Complying with the directives is a particular challenge for banks, exposed to a three-fold risk in terms of money laundering activities.
First of all, there are powerful sanctions imposed by supervisory authorities for unsuccessful compliance with anti-money laundering obligations.
According to experts estimations, between 2008-2018, inefficient management of AML and KYC areas have cost banks a total of $26bn.
Secondly, a financial institution whose system will be used to legalise funds derived from criminal acts, is not perceived in the public opinion as a victim, but rather as a partner in crime – or, at best, as an incompetent entity that was not able to properly secure its infrastructure.
Finally, along with the development of cybercrime, frauds and scams committed to the detriment of financial institutions and their clients have an increasing share in the perpetrators' revenues.
AI algorithms taking a lead
In order to improve this situation, AI algorithms should take a lead. The biggest benefit of using advanced analytics and artificial intelligence methods is a more accurate identification of suspicious cases and a significant reduction of false alarm levels.
AI in AML optimises the existing processes by significantly enhancing the effectiveness of inefficient rule-based approaches, characterised by high false-positive rates, and unable to consider complex interdependencies between various activities carried out to launder money.
Experts have been observing the increased interest in AML solutions using artificial intelligence on the US market for some time. This is mainly due to the position of local regulators who encourage financial institutions to use innovative approaches in counteracting money laundering and combating financial crime.
Europe lags behind
In contrast, European financial institutions seem to be not that happy when it comes to successful AI implementations. Although there are some banks in Europe that develop their AML systems based on modern methods, including machine learning, the traditional customer segmentation and manual monitoring of fraud alerts still prevail. This is the result of a lack of universal communication between the financial sector and regulators.
There have been many cases when representatives of financial institutions feared the implementation of artificial intelligence to improve the AML process because they didn’t know how regulators would react. Despite the fact that this problem mostly affects CEE, Western Europeans are still lagging behind compared to Americans.
Open communication is key
These days, there is no doubt that this situation leads nowhere. Without a significant transformation in this area, meeting AML obligations will sooner or later encounter serious obstacles, where classic solutions might be completely helpless. In order to speed up the AI-based transformation process and avoid unnecessary problems, it seems crucial to arrange a deep and substantive dialogue between regulators and representatives of financial institutions, perhaps at European level. This attitude would allow full transparency of the process, open communication between its key entities and extraordinary value for the European economy.