Fighting financial crimes like fraud and money laundering is getting harder for law enforcement and banking executives in an increasingly digitalized environment. Criminals continue to stay up with technological advancements and use ever-more-advanced tactics.
Despite being necessary and advantageous for law-abiding consumers, data privacy rules restrict financial institutions’ capacity to conduct routine commercial activities. In addition, banks are under pressure to continue operating efficiently while maintaining regulatory compliance.
The International Monetary Fund claims that money laundering incidents have even led to bank failures and damage public confidence. The quantity of money laundered can be a good indicator of the influence on public trust, albeit it can be difficult to measure exactly.
For instance, three money laundering networks moved $670 million through TD Bank accounts thanks to TD Bank’s anti-money laundering measures’ shortcomings. Even with the Financial Action Task Force’s efforts to provide a clear framework for workable anti-money laundering measures, banks continue to encounter major challenges in their efforts to apprehend transnational criminals.
On the “AI in Business” podcast, Senior Editor Matthew DeMello spoke with Standard Chartered Bank’s Nick Lewis about the ongoing evolution of anti-financial crime initiatives.
Three main conclusions from the discussion will be the subject of the essay that follows:
Using deterministic AI systems to detect financial crime: Recognizing the need for human input to overcome the shortcomings of basic, rules-based AI systems in identifying irregularities in financial transactions and differentiating between lawful and illegal activity.
In order to overcome restrictions relating to confidentiality issues, financial crime investigations must emphasize the need of human judgment while balancing AI automation with human monitoring.
Improving information exchange to combat international financial crime: pointing out that in order to fight money laundering networks, banks and law enforcement must share data more quickly and across borders.
Nick Lewis, Managing Director of Standard Chartered Bank’s High Risk Client Unit, is the guest.
Proficiency in Threat Mitigation, Risk Assessment, and Leadership
Brief Recognition: Nick led multiple noteworthy operations during his more than 30 years of service in UK law enforcement. From 2008 to 2013, he worked in the British Embassy in Washington, DC, as a Counselor for Transnational Organized Crime. He received the National Intelligence Medallion from the US government in March 2013 and an OBE from Her Majesty the Queen in 2014 for “services to international law and order.”
Using Deterministic AI Systems To Identify Financial Crime
Lewis makes it clear from away that crime is not a theoretical idea. He characterizes crime as a sequence of actions that are predictable in addition to being extremely traceable and mappable. He notes that a lot of offenders tend to follow a repeating pattern of actions because they believe that if something worked for them yesterday, they will probably do the same thing the next time.
According to him, the financial industry has historically addressed that by creating red flags, which identify recurring patterns of behavior and then adjust them for rules-based systems. Red flags are generated based on the establishment of individual regulations. Based on such red flags, companies then generate alerts and look into them to see if they are suspicious.
Lewis warns that criminal activity is not always indicated by red flag behavior. He believes that using AI to map typical consumer behavior patterns and evaluate human behavior features can help identify whether a behavior is abnormal within a range of norms. This will help start to differentiate between legitimate and problematic activities.
An individual with a single employer who suddenly receives payments into their account from several sources, marked as salary, is an example of abnormal behavior that AI may be able to detect. Lewis says that businesses may better comprehend the context of those warnings by integrating them with publicly accessible data from sources like a person’s LinkedIn profile. Lewis asserts that context is crucial in financial crime.
“A transaction is a snapshot of a second of one dimension of one element of the client’s behavior on that day,” Lewis concludes clearly.
Banking: Finding A Balance Between AI Automation And Human Oversight
When questioned about how banks can effectively balance their investments in maintaining and investing in human monitoring to determine when to assist law enforcement with the development of AI systems to detect anomalous banking activities, Lewis replies that he sees it as a three-step process:
- Step 1: Establish a basis for detecting anomalies by using analog-like, legacy rules-based techniques to discover departures from established norms and behaviors.
- Step 2: Use machine learning to find warnings that don’t seem to be related to financial crime
- Step 3: Use AI to forecast risks by combining various data sources and offering context for analyzing the client’s actions and the transaction.
Lewis continues by outlining the two primary reasons why the banking industry is concerned about AI. First, AI tools usually work best in open-source settings; however, banks are unable to import large volumes of external data into their systems or expose sensitive customer data to these settings because of confidentiality regulations.
These restrictions call for a cautious approach to AI implementations, one that removes unnecessary Internet noise while protecting customer data. Second, even though AI can automate manual operations and cut down on the need for workers, it still lacks the human ability to make decisions, especially in cases involving complicated financial crimes.
Lewis says the problem with AI is that it doesn’t like not being able to answer, so it will compile all the data it has and present what it thinks is the best response, but it won’t accept uncertainty. Lewis warns that companies must exercise caution and make sure they aren’t utilizing AI to make conclusions that are different from what people would decide because of the “false confidence” that comes from AI interactions. Lewis therefore believes that human investigators will always be required, especially in instances that are extremely complicated and nuanced.
Improving Information Exchange To Combat International Financial Crime
Lewis disputes the widely held belief that law enforcement has an edge when looking into financial crimes. He emphasizes that, in contrast to banks and law enforcement, criminals have a significant edge over law enforcement because they do not respect borders and are consequently unhindered by national or jurisdictional restrictions.
Internal data sharing between nations is restricted, even for big international banks. Because of this, banks find it very difficult to provide any one law enforcement agency with a comprehensive picture of global money laundering networks.
He tells to listeners, “For large international organizations like ours, we are also hindered and constrained internally by our ability to share information and data across borders, even within our own institutions, let alone share information with another institution.”
Despite banks’ mandatory duties to assist law enforcement, especially with relation to money laundering networks, these challenges nevertheless exist:
“I am unable to identify a single law enforcement agency. I am unable to identify a single law enforcement organization. I have to dismantle the entire network and inform the police or FIUs in each of those nations. The part that pertains to them is what I must tell them. I’m sometimes even forbidden from informing them that there is a dimension in a different nation.
– Nick Lewis, Managing Director of Standard Chartered Bank’s High Risk Client Unit
According to Lewis, in order to effectively combat global financial crime, leaders in both industry and law enforcement must support mechanisms that facilitate faster and easier cross-border information sharing. This will improve the financial institutions’ pronounced disadvantage when compared to the criminal enterprises they are protecting themselves against.