Talking Financial Crime Prevention with Experts
In early June, the Financial Times presented “Using the Power of AI to Prevent Financial Crime” as part of the FT Digital Dialogues. I thoroughly enjoyed being part of the panel conversation with a group that included one representative each from Technology (me), Banking (Natasha Meaney, Global Head of Financial Crime Prevention and Conduct Risk for UBS) and Regulation (Tom Neylan, Head of Policy Development in the Secretariat of the Financial Action Task Force, or FATF). When I’m asked to speak, I’m always aware that I work for a company that makes money from products that are based on some combination of software and advanced analytics, and that there’s a healthy skepticism on the part of any audience when I describe the benefits of the solutions RDC provides. I was delighted that the three of us seemed to be very much in sync with one another about the criticality of using tools like AI Review to fight financial crime.
Financial crime is a huge problem that’s only getting bigger – financial crime represents 5% of the world’s GDP. And collectively we’ve barely made a dent in the fight to prevent; only a very small percentage of financial crimes are caught. So we have to use every tool at our disposal that proves effective, while ensuring that the solution doesn’t itself become a problem. After all, “the treatment worked, but the patient died” is not a good result. Predictive analytics is one such tool. Natasha put it very well when she pointed out that AI solutions reduce the noise, allowing people to focus on the important things (she mentioned efficiency gains of 80-90 percent). Tom and Natasha both highlighted the importance of access to comprehensive customer and transaction information, while maintaining privacy, and the potential that could be unlocked by allowing financial institutions to pool information. I’m aware of the privacy, legal and policy constraints in place that make it difficult for banks in many jurisdictions from sharing information. But the analyst in me becomes really excited when I think about what we all could do – banks, regulators and even vendors like RDC – if that information could be gathered together and analyzed to catch criminals.
I was interested in the differences in perspective and experience when Matthew Vincent, our moderator, asked about AI being used to fight coronavirus-related financial crime risks. Tom described the tension between regulatory oversight and the need to get relief out quickly to individuals and businesses. Natasha emphasized that fraudsters are behaving differently, and technology must adapt. She added that the sheer volume of transactions has increased, and coupled with changes in individual behavior, making it necessary for us to adapt the way we try to identify and catch criminals. I encouraged anti-fraud and anti-money-laundering technology and service providers to offer their products at no charge to hospitals to fight against fraudulent suppliers and the Small Business Administration in the US to help them ensure that relief money goes to legitimate businesses and not criminals.
When asked what the most important point we’d like to see implemented, Natasha said the way to tackle financial crime is to use a suite of tools, including humans, and not just AI. I couldn’t agree more. Tom shared that we have huge potential to fight financial crime, but we have to connect up organizations that currently aren’t speaking to each other much. My final word was that while ML is powerful, it is not truly “intelligent.” When you’re thinking of adopting an AI-based tool, make sure you understand the limitations, and if you can’t find someone who can explain it in terms that you can understand, then they probably don’t know what they’re talking about. We all agreed that the combination of the insights provided by AI and ML-based tools with human guidance and sharing would meaningfully benefit us all in our fight against financial crime.
Here’s part of our conversation: