The FinCen regulations based on the Anti-Money Laundering Act (AMLA) of 2020 contain a number of updated AML obligations, requirements, and guidelines applicable to financial institutions, including banks, investment structures, and others. Among other issues, the AMLA focuses on the technological aspects of combating money laundering.
For the purpose of complying with new and updated regulations, financial institutions (FI’s) implement a variety of technological systems for transactions and customer monitoring. Such systems scan all transactions, along with the businesses and individuals involved, for irregularities and suspicious behavior. Identifying companies and individuals who are listed on various sanctions lists is one of the most critical aspects of preventing money laundering. Unfortunately, this task poses significant challenges. There are hundreds of thousands of names in different languages, as well as double entries, spelling variations, and varied names. A PARTICULAR challenge is dealing with structured names, long names containing a lot of alternative spellings, abbreviations, etc.
For the sake of illustration, consider a Chinese name:
艾美 刘 周
and its possible English variations:
- Amy Li Zhou
- Amy L. Chow
- A.L. Zhou
- Amily Show
- Amy Lii Zhou… the list of possibilities is endless.
Which system is capable of cross-checking this name throughout several databases and determining they are the same?
Currently, there exists a wide range of rule-based and AI-based technological platforms that are aimed at and used by banks and other financial institutions to solve this problem. They are, however, utterly unsuccessful. Either they flag all entries containing Amy, resulting in a long list of False Positives, or they fail to connect two entries that are not 100% identical, resulting in False Negatives.
Legacy rule-based systems are inefficient because they are limited to a specific set of rules (conditions) that cannot take into account all variables, errors, and spelling variations associated with a given name. Recently, AI and machine learning-based systems have made their way onto the market. With the growth in market demand, there is also a growth in the number of new technological tools that are intended for monitoring transactions as well as flagging the transactions, businesses, and individuals that trigger the alert.
Artificial intelligence-based systems are not as effective as they might appear, either. They produce just as many false positives, making it impossible to identify suspicious activity or problematic transactions.
Additionally, rule-based and artificial intelligence-based systems are inherently biased, raising, in addition to social and moral concerns, the question of non-compliance with the recently updated anti-discrimination regulations.
In practice, however, even those financial institutions investing millions of dollars, euros, or pounds in technological solutions fail to properly monitor and identify threats, and as a result, suffer huge financial losses resulting from fines and considerable damage to their reputation.
Consider, for example, the recent NatWest money laundering case. The NatWest bank failed to comply with FCA regulations and broke UK anti-money laundering laws. The Bank claims it has invested £700m in transaction monitoring, customer screening, and due diligence technology. Further, it employs over 5,000 staff members in its anti-financial crime and compliance departments. This did not help. A large number of false positives caused the employees to disable one of the monitoring and alerting rules of the system…
Just this year, Capital One was fined $290m for failing to comply with Anti-Money Laundering regulations in the US. It is indicated that Capital One “often failed to detect and report” unlawful transactions.
We cannot even claim that this problem is new and that FI’s are simply adapting to the new rules and regulations. Wachovia Bank was accused in 2008 of failing to identify and prevent money laundering by drug cartels and Hezbollah. The company avoided prosecution by paying a $160 million fine and promising to enhance its anti-money laundering procedures.
In 2004, Standard Chartered Bank had “some problems” due to the absence of anti-money laundering systems. In 2012, the Bank paid a fine of $670 million for violating sanctions against Iran (2005-2006). As well, the Bank has been accused of violating sanctions against Burma, Libya, and Sudan. Standard Chartered Bank was fined $1.1 billion for failing to implement AML practices.
Danske Bank, Nauru, BCCI… and more.
Simply put by The Economist, “The global system for financial crime is hugely expensive and largely ineffective.” They go on to say that “the war against money-laundering is being lost.”
The time has now come to find a truly efficient solution for this war before it is lost for good. Such a solution could prove pivotal on the way to victory.
FinCom has developed a revolutionary phonetic fingerprint solution for name screening that is not only promising but has already established itself as one of the most effective and efficient monitor/alert tools available on the market today.
Based on advanced phonetics, sophisticated mathematical algorithms, and computational linguistics, FinCom’s patented technological platform:
- Ensures the integrity of the information in the databases
- Formats and transforms data into a mathematical code using over 44 algorithms (phonetic fingerprint)
- Matches entries between databases using in 40 languages accurately and efficiently.