Artificial Intelligence (AI) Catches Up with Financial Crime
An AI approach to anti-money laundering does not require rules to identify potentially criminal transactions. Instead, algorithms power the engines that runs AI and machine learning technology. These models operate on large data sets to process patterns, anomalies faster and more efficiently than a human.
Machine learning creates and relates risk profiles across a vast data network to determine whether a bank should continue doing business with a potential client (or not). The system would be trained to identify suspicious transactions over time and identify patterns and connections such as where a customer opens an account relative to their home address or even changes in the customer’s social media presence.
However, considering the breadth of the latest in artificial intelligence advancements such as robotic process automation (RPA), natural processing languages, and graph analytics, may be a mind-boggling affair for any business looking to augment their anti-money laundering technology. There needs to be a balance of enthusiasm in applying these technologies and limiting exposure to unnecessary risks.
To put it simply, it should not automation for the sake of it. The focus of any anti-money laundering team should not be on investing in a machine learning model that cuts down false positives by 40% but to understand the inner workings of the algorithms powering the intelligence.