WITH cross-border communications soaring and digital transactions becoming more rampant, fighting money laundering has become more complicated than it previously was.
As a result, financial institutions are experimenting with all kinds of technology to safeguard their customers and block out bad actors. Increasingly, in the fight against money laundering, artificial intelligence (AI) and more specifically, machine learning (ML), seems to be providing a strong defense.
Whether you look at international giants such as HSBC and Standard Chartered Bank or regional banks such as Maybank and DBS, everyone is collaborating with fintechs or developing their own in-house AI/ML solution to fight financial crime.
Banks, due to their very nature, have access to tremendous volumes of data. Hence, with AI/ML on their side, they’re able to quickly build models and frameworks that learn what “ordinary and regular” looks like and pinpoint transactions that should be scrutinised further.
The number of non-cash transactions in Asia is expected to reach 277 billion by 2020, according to the World Payments Report 2017 — and there’s no telling which of those are being initiated by bad actors intending to launder money using new and innovative schemes.
Being automated, for the most part, it helps make executives much more efficient and allows the bank or financial institution to pay attention to every single transaction despite the large volumes.
EY, for example, helped European financial services giant Nordea build a model that helped weed out false positives and put the spotlight on transactions that actually deserved immediate attention.
“Like the algorithms that help online retailers target customers according to their taste in brands, the model we built with Nordea became more insightful and accurate as it learned how to interpret the signals that indicate criminal activity, analysing vastly greater quantities of data, more reliably, than existing people or processes could,” explained EY executives through a case study.
Why use AI to fight financial crime?
“By using such techniques and the power of automation, compliance teams can investigate true compliance risks and suspicious cases, reduce manual, and repetitive tasks and perform higher value work such as analysing the outcomes and quality to ensure policies and procedures are substantial,” Deloitte Southeast Asia Financial Crime Compliance Leader Radish Singh told Tech Wire Asia.
“With technology used more effectively and innovatively, humans can focus on the material risks.”
According to Singh, with digital disruption, technological revolution, open APIs, and emerging typologies, there is a widening gap between where a financial institution’s financial crime compliance programme is, versus where it needs to be.
PwC recently studied the AI readiness of financial institutions with respect to financial crime and found that part of the problem is that business leaders don’t really know where to start, or conversely, are thinking too far ahead to techniques that are feasible but not practical yet.
In fact, PwC’s consultants believe that most financial institutions would be better served if they focused on what’s possible over the next one to two years, rather than what AI could do five years down the line.
In the short term, experts believe that AI could be used to augment processes and track transactions that escape set benchmarks — when observed in isolation or when seen in combination with other related transactions — aiding compliance professionals within the financial institution.
“When it comes to business efficiency, using new technologies such as machine learning has tremendous potential to reduce manual processes, and even streamline repetitive tasks that often weigh compliance and operations teams down.
“This alleviates the cost burden and makes compliance a more meaningful exercise. In addition, machine learning can improve a financial institution’s accuracy in detecting money laundering and terrorist financing risks,” said Deloitte’s Singh.