Whenever a new technology enters a market it comes with both tales of hyperbolic virtue and the need for extreme caution. The introduction of AI and ML tools to fight financial crimes has been no different. According to Carl Case, Principal at EY and a leader in EY’s Financial Crimes Compliance and Technology Practice, there are three myths that AML professionals need to dispel quickly so they can begin to take advantage of the explosion of AI and ML in financial crimes monitoring and detection technology.
Myth 1: AI is black box technology and won’t yield different results.
If there is one myth that Case could dispel it would be that AI is just another black box technology that won’t yield different results from rules-based systems that are in use at financial institutions today. “I understand where this opinion comes from,” said Case. “AML teams are used to working with proprietary algorithms, but AI and ML have experienced a recent, great leap forward in the transparency of models and their ability to provide explainable rationale. Combined with continually improving algorithms to address detection efficiency and efficacy and you have a technology that can make a material difference in the fight against financial crime.
Myth 2: A team of data scientists is all that is needed to get an AI solution up and running.
“This is simply not true,” says Case. While you can certainly take that approach, financial institutions would be better served working with their AML team and technologies that improve data quality to get the solution up and running using their domain expertise and understanding of the business problem. “Machines learn based on the data that you feed them, and without involvement from the AML team, data scientists can sometimes fail to recognize where that learning has gone awry or on the flip side, fail to recognize success, getting trapped in analysis paralysis because they lack specific subject matter knowledge. It’s another instance where perfect is the enemy of good.
Myth 3: Regulators will penalize my institution for using new technology like AI and ML.
Case says that it is more likely that regulators will begin to penalize financial institutions for not using the best technology available if it would have helped detect crime, a sanctions violation, or other malfeasance. “The US Treasury’s Financial Crimes Enforcement Network recently released a Joint Statement on Innovative Efforts to Combat Money Laundering – in which regulators encourage the responsible adoption of AI and ML to improve the ability of institutions to detect financial crime,” he said. U.S. regulators have previously levied fines for failure to detect suspicious patterns of transactions and the fines are getting more significant and impactful.