Although the dire need to put a stop to money laundering is well understood, an ongoing negotiation regarding what amount of money laundering is an acceptable cost of doing business plagues regulators and those inside who are responsible for the detection and prevention of financial crimes.
Should we judge those tasked with stopping financial crimes for possibly allowing an acceptable number of money laundering transactions? Unfortunately, the teams in financial institutions in charge of thwarting financial crimes are hamstrung by limited time and outdated technological resources.
Change however, is happening now. Over the last several years, advancements in data science, including artificial intelligence (AI), machine learning and big data management, promise to stifle money laundering making the accurate evaluation of all transactions a viable reality.
The utilization of AI and machine learning dramatically improves the effectiveness of money laundering investigations, providing the with the ability to efficiently identify all money laundering transactions, regardless of the dollar amount.
The technology can improve "acceptable costs." Financial institutions and issuers set dollar, volume and velocity thresholds for transaction monitoring. Above these thresholds, they attempt to screen all transactions for money laundering. Below them, they accept that some money laundering, terrorist financing, and other financial crimes might go unaddressed.
This Above the Line, Below the Line (ATLBTL) practice functions on a risk-based transaction monitoring process by which not all transactions are screened for possible money laundering and other financial crimes concerns.
The percentage allowed to pass unscreened is determined by institutions and federal regulators that work with them to ensure that they comply with AML regulatory guidelines. Industry estimates state these allowances run between three and 10 percent of all “below the line” (BTL) transactions. Institutions and regulators conduct periodic assessments to ascertain if the ATLBTL thresholds are functioning within their established risk tolerances.
Regulations on this topic continue to be rolled out, at the state, national and global levels. These include
Given technological advancements, ATLBTL is one example of the many compromises financial crimes teams no longer should have to make. AI for AML is built for the purpose of complementing transaction monitoring systems, increasing investigative efficiency, driving out false positives, and catching the false negatives that embody financial and reputational risk. With trillions in laundered money still going through banking systems, AI offers the opportunity to move toward a new standard.
