Bringing true AI to AML

Stephen Moody

Criminals have become increasingly sophisticated at money laundering, creating a flourishing industry that costs banks around the world a great deal of money in fines each year. Illegal funds flow in ever-greater volumes through the global financial system; fraud is the most common crime in the digital world, and criminal capabilities outpace the industry’s ability to respond. 

In parallel, anti-money laundering solutions are also becoming a bigger business, as financial institutions face increased penalties and fines for control failures. A positive is that artificial intelligence and machine learning technologies offer a great deal of promise in terms of combatting money laundering. However, a lot of the so-called AI/ML solutions out there aren’t using true AI — it’s just window dressing on old, outdated approaches. Financial institutions need a solution that truly harnesses the ability and power of ML/AI to combat money laundering. And with the Anti-Money Laundering Act of 2020 passed in the U.S., there’s an increased imperative. These new regulations are considered the most consequential anti-money laundering legislation passed by Congress in decades. It includes increased penalties for violations, enhanced protections for whistleblowers and increased authority for U.S. regulators to seek documents from foreign financial institutions, among other provisions.

As money laundering booms, regulations increase

Money laundering has proliferated yet further during the pandemic and its resulting shift to working from home changed the cybercrime landscape and created new opportunities for money launderers and financial criminals. Financial services organizations were forced to conduct more of their business online, and the new global home workforce created additional opportunities to exploit financial and enterprise cybersecurity weaknesses. The combination of governmental financial support and increased online and remote banking has led to a spike in financial fraud and associated money laundering.

Enforcement has changed during this time, as well. Non-compliance is a costly option as fines levied against banks for AML violations continue to increase. U.S. banks were hit by $14.2 billion in fines in 2020, the bulk of which was for AML violations. The passage of the U.S. Anti-Money Laundering Act of 2020 ushers in a new era of anti-money laundering enforcement and regulation for financial institutions in the U.S. With these new rules in place, including the potential for enhanced sanctions and fines, financial institutions have a new imperative to take AML seriously.

Financial institutions are spending a great deal of money on AML solutions, but the problem persists. A study by Lexis Nexis Risk Solutions found that in 2019, AML compliance across U.S. and Canadian financial services firms cost $31.5 billion. The study found that a layered approach is more successful — especially ones that entail ML/AI — but that many firms are still relying on manual efforts when it comes to AML compliance technology, which isn’t optimal for performance or cost.

Anti-money laundering systems are overwhelmed by false positive alerts, forcing banks to employ armies of investigators who spend 99 percent of their time looking at completely normal financial behavior. Fortunately, the AI revolution can help with this challenge. 

The true AI difference

Many of the AI/ML solutions currently on the market rely on a technique called supervised machine learning. Essentially, the AI is given a history of things that are true, and then they try to find new examples. A classic example is cat recognition. You show the AI many photos of cats, and you show it photos that are not cats, and then it learns — so it can identify a photo as “cat” or “not cat.” But in real-life situations involving crime, it’s not so cut-and-dry when it comes to identifying what is crime and what is not. 

Criminals are good at hiding, so what’s needed is a more sophisticated type of AI that can enable mapping of behaviors within a system. Let’s say you work for a bank with 10 million customers, and they’re all conducting thousands of transactions per year. So, for instance, Customer A sends money to Customer B, and Customer B sends money to Customer C. There are many different types of transactions in different amounts. Somewhere hidden in all that, there is money laundering that no one can find. It’s just not possible to know what you are looking for in advance.

However, using the right blend of techniques, you can start to find the hidden areas of crime. A key step is to first learn and map all of the different types of behaviors associated with your different customers. ML technology is available today that can learn money flow and financial behavior and compare all the different customers’ behaviors to each other. This reveals pockets of higher risk behavior in a natural way. The banks’ historic view of risk (based on investigation outcomes) can be overlaid on this map enabling the system to focus on particular areas and generate alerts. This kind of automated solution supports regulatory review and achieves a high level of performance in accuracy and risk coverage.

Getting the upper hand

Both money laundering and the anti-money laundering industry are big businesses. As more AML solutions proliferate, AI and machine learning are playing a larger role, but the reality is that you can’t just throw standard machine learning techniques at the problem and magically solve it. Not all AI is true AI; payments and banking professionals must carefully vet AI-based AML solutions to find those that can automatically map behaviors. This saves human resources and increases accuracy, decreasing risk and fines. As criminals increase their sophistication at hiding their activities, AI has become a necessary crime-fighting partner.


Stephen Moody is the chief innovation officer at Symphony AyasdiAI. He brings over 15 years of industry experience developing market-leading solutions to combat financial crime. Previously with Simility, ThreatMetrix and BAE Systems, he has led teams that have pioneered advanced fraud and financial crime solutions for organizations in financial services, commerce, telecommunications and government. He has a Ph.D. in astrophysics from Cambridge University.