Bringing AI into the fight against money laundering

Money laundering is now the world’s third-largest ‘industry’, with the UN’s drugs and crime office estimating that over $2trn is laundered every year. The traditional gambling industry has in the past offered plenty of opportunities for such activities, particularly while it was mostly cash-based.

However, egaming represents a different animal. In mature, regulated markets, money laundering isn’t easy, consumer tracking is more intelligent, and security is high.

While the advent of egaming, alongside more stringent regulations globally, has made money laundering trickier, significant challenges remain. AML initiatives now range from tackling criminal leisure spend to ‘smurfing’, which involves the distribution of cash into smaller transactions to evade threshold requirements and therefore suspicion.

Until recently, the industry has tackled the identification of crime online primarily by using knowledge-based mechanisms such as rules-based systems that embed basic regulatory requirements within them.
While capable of reflecting requirements focused on simple thresholds, these systems are often found lacking. They are, for example, unable to proactively monitor the activity of millions of online customers and evolving criminal activity behavioural patterns that fall above or below regulatory thresholds.

Upping their game

These systems must change, and the use of AI can drive that shift. If deployed correctly, the use of AI will raise AML standards. BetBuddy, Playtech’s responsible gambling data analytics business, has just partnered with City, University of London and egaming operator Kindred Group to explore the use of AI in AML initiatives.

It’s a three-year-project but the early signs are positive. Some of the techniques available today to detect anomalous behaviour, including state-of-the-art recurrent neural networks, have shown promising results in the analysis of complex data. Broadly, AI can help to assess patterns, trends and anomalies, and even correlate networks IP addresses with crimes. Anomaly detection techniques may be useful at flagging suspicious behaviours that traditional threshold-based approaches would never be able to flag.

Other data sharing initiatives, such as the open banking APIs, will provide new ways for customers to securely share personal data about their banking and to help address affordability initiatives. Of course, there are barriers to its success; AI often requires large investments and long-term outlooks, and only the largest tech companies can afford this luxury. Egaming companies that do have AI strategies are being driven by an immediate business need. But it’s difficult to attract talent, and develop a strategy, when such a laser focus on a business goal reduces the scope for innovation and creativity, both are crucial in AI.

We need a culture shift. Revenue-generating teams need to see the value in class-leading compliance capabilities because it can be a commercial advantage in preventing regulatory action. This should extend to investment in AI, which itself should be viewed as a vital tool in compliance. A bonus for operators is that AI can also be used in both responsible gambling and product personalisation.

The fight against money laundering will always go on, but if operators shift their thinking and start using technologies like AI they can take a step towards winning it.

Simo Dragicevic is the founder and CEO of BetBuddy, a specialist behavioural identification and modification platform developed for regulated gaming markets. A veteran of the financial services industry, he is the external supervisor for City, the University of London’s Research Centre for Machine Learning Ph.D programme. BetBuddy was acquired by Playtech in 2017.