A large portion of financial crime is going undetected due to data problems causing high levels of false positive rates in financial institutions’ fraud detection systems, according to Stuart Breslow, managing director of technology and policy at Google.
“The error rate was very gently put as significant,” said Berslow on a panel at the Sibos conference in London this week. “The error of the false positive rate in these systems currently verges on just about 100 percent. The vast majority of G-alerts that come through the systems that are currently used, are all false positives.
“If you are an investigator knowing that every day you are walking in, and every day you are clearing virtually all of the alerts that you are executing on, that is really not an inspiration to get out bed in the morning.”
According to a report from Chartis Research and Accuity on financial crime compliance screening, 66.1 percent of respondents - of which 83 percent were financial services firms - said they face significant challenges in achieving low false positive rates.
For Breslow, this large false positive rate means “very little actual financial crime is being detected” as smart criminals are aware of the systems that are currently being used.
“All the network analysis within a single bank is of great value, but the reality is, the purpose of money laundering is to move money from institution A, to institution B, to institution C. So, you are really not necessarily going to see something within a single institution that is going to tell that fact is the money laundering,” he said.
“The other thing you need to think about broadly is what is the way in which institutions will share information about their common customers or about the money that goes through those institutions to better detect what is really a global scurge and the amount of money laundering going on in financial services right now.”
For Marc Fungard, global head of research and analytics at HSBC a lot of financial institutions are struggling to scale fraud prevention systems.
“When you go looking for financial crime… it is a very fringe phenomenon in a very wide market and so it’s having to differentiate that which you need to find, from that which you don’t,” said Funguard.
“We know from talking in the industry the false positive rates around trying to do that are not where they could be or where they should be which gets us to a question of how do you actually differentiate at the scale and size of an HSBC or some of the other institutions in the room and begin to understand what it is you need to know.”