If we take a step back and look at the state of financial crime, it’s a worrying picture. Increasing fines from regulators are a stark reminder that criminals continue to outsmart even the biggest financial institutions. In 2020, more than $2 billion in global money laundering fines were issued – five times the amount issued the previous year. What’s more, authorities levied almost $1 billion in AML fines in 17 big actions in the first half of 2021 alone. And this shows no signs of slowing down.
It’s clear that transaction monitoring cannot cope in its current state. So, how can we use emerging tech like artificial intelligence (AI) to solve the issue?
The problem with legacy tech
Tighter regulation and implementing fines like the above are helping to address shortcomings in AML management, monitoring of suspicious activity, and customer due diligence. The EU, for example, has announced the introduction of a new AML authority for greater coordination. However, tighter regulation in itself is not enough.
Crucially, technology is evolving at an incredible pace, meaning so are the tactics of criminals. Techniques for carrying out financial crime are evolving all the time – they aren’t standing still, and banks shouldn’t be either. Many of the tools currently used by incumbent banks are anchored by legacy tech and rely too much on human intervention. They simply can’t keep up with the increasing sophistication of criminals.
Digitalisation, therefore, is key. The success of AML can only improve when banks fully embrace the technology revolution, such as implementing AI.
Work smarter, not harder
The traditional rule-based approaches to transaction monitoring are outdated, using static thresholds that only capture one element of a transaction. Because of this, they deliver a staggering number false positives and its well known the industry sees false positive rates of between 97 and 99 percent. It’s incredibly inefficient.
What’s needed is an AI-driven model that can consolidate multiple risk factors at once. AI can assess a number of dimensions on a transaction, extracting a risk score, and developing an intelligent understanding of what risky behaviour broadly looks like. By looking at a broad array of factors that come together in a payment. The result is something more akin to how money laundering works in the real world. Not only does this reduce false positives, but it also means fewer requests for information on payments – key for an improved customer experience.
And the good thing about AI is, it’s constantly learning with the more data it is fed. With every analysis it becomes a more intelligent solution.
The proof is in the pudding
As a licenced payments bank, Banking Circle is pioneering the use of AI in AML but that doesn’t mean there hasn’t been some resistance from the industry. By regulators in particular, AI is often seen as unproven and risky to use – but the opposite is true. Since Banking Circle’s initial implementation of its AI model three years ago, the improvements to alert quality have been incredible. AI has alleviated admin-heavy processes, enhancing detection by increasing rules precision and highlighting red flags the naked human eye could never spot.
In fact, between 2019 and 2021, Banking Circle reduced the amount of false positive rates exponentially. More than 600 accounts were closed or escalated to compliance due to AI-related findings – an increase of 380 percent. What’s more, as payments rose by 150 percent, the number of alerts generated fell by 30 percent.
A holistic view of AI in AML
Results like Banking Circle’s show just how effective AI can be in the fight against financial crime, but to ensure wider adoption, it’s crucial that regulators and financial institutions see the role of AI in AML within the broader framework of digital transformation. For instance, it should be made obvious that it is not meant to replace humans with machines, but empower them.AI will enhance efficiency, and free up resources for employees to focus on other value-adding areas, such as customer relationships.
Attitudes are beginning to change, particularly in the payments and FinTech space, but there is still some way to go. With the United Nation’s estimate that the total amount of money laundered annually is anywhere between $800 billion – $2 trillion (or 2-5 per cent of global GDP), there has never been a better time to make a change and embrace AI for AML.
