In my former life as a bank attorney-turned career money launderer, one of my priorities in the Placement phase of operating a laundering pipeline was the transnational transfer of criminal proceeds abroad. Whether it is the repatriation of narcotics proceeds overseas, back to kingpins, safe storage of the profits of white collar crime in offshore tax havens, or the first stop in the traditional wash-dry-fold operation (Placement, Layering & Integration), one must generally first move that dirty money somewhere, rapidly and without being detected by the financial services industry.
The task of evolving transaction software has always been to shine the proverbial light on those transfers where previous platforms have failed, a goal that in the past has not been achieved. This has become more critical as regulatory and law enforcement agencies have sought to make the financial services sector personally responsible for their AML failures, and this exposure has, at times, even descended to compliance officers themselves, for what some agencies now regard as negligence, or even compliance malpractice.
We know how difficult the task of identifying outbound funds transfers are, when they occur, randomly, sporadically or confusingly, within otherwise normal and legitimate International payments in the ordinary course of business. Money launderers know about that vulnerability; they insert themselves within legitimate payment systems, through creating corruption, through paying substantial sums to become silent partners, or even majority owners of legitimate businesses, and through whatever dirty means they can employ so that they can move millions of dollars right through your bank with impunity, month after month, until they have achieved their laundering goals, and move elsewhere, leaving a financial institution to explain its inability to find and report financial crime in real-time, with all the negative press that goes with it.
The employment of advanced machine learning and statistics, within a system using Artificial Intelligence, results in the detection of financial crime that previous rules-based platforms could not ferret out. Patterns in raw transactions formerly too opaque to be recognized and classified, are not only identified, they are generated for the user, notwithstanding the volume being examined. The algorithms first identify, then group the data, find patterns, and ultimately classify the information, to discover any hidden structure.
If I was still a practicing money launderer, I might be looking to get into another line of work; next-generation machine learning programs, powered by AI, have the ability to find those transfers. As soon as the financial services sector begins acquiring those systems, there will be a sea change in AML abilities in transnational transactions.
For further reading: Exposing Financial Crime