AI is supposed to detect changed money laundering methods

When entire economies around the world shut down due to the Covid-19 pandemic earlier this year, it wasn’t just entrepreneurs and consumers who had to adapt. Criminals too suddenly had a problem: how do they launder their money? “The money is still coming in, but there is nowhere to go,” said Isabella Chase, who works on financial crime at the Royal United Services Institute, a UK-based defense and security think tank.

After all, the usual ways of passing on profits from organized crime through legitimate corporations – often across borders – until there was no clear path back to their source, were barely accessible. Many of the best money laundering companies have been hardest hit by the pandemic: small shops, restaurants, bars and clubs are because of the high volume of cash in circulation. Large cash deposits became difficult when bank branches and remittance services like Western Union closed.

More from Technology Review

According to the United Nations Office on Drugs and Crime, between two and five percent of global gross domestic product – currently between $ 800 billion and $ 2 trillion – is lost each year. washed. Most cases go undetected. It is estimated that only around one percent of the profits generated by criminals are confiscated. But that was before the pandemic. Since then, fraud has increased and fears about Covid-19 created a lucrative market for counterfeit protective equipment and medicines. More people were spending far more time online, creating a larger pool for phishing attacks and others Cases of fraud. In addition, drugs are still bought and sold.

The lockdown only made it more difficult to hide the proceeds from these criminal deals at the beginning. When the normal channels for money laundering were closed, new ones opened up at the same time. Thanks to government bailouts, huge sums of money are once again flowing into small businesses. This led to a flood of financial activity that obscured money laundering.

The pandemic also forced the anti-money laundering teams to break new ground and look out for abnormal activities. New artificial intelligence tools (AI) are proving to be the most important weapon. The systems of many larger and older financial institutions are still based on hand-crafted rules, such as that only transactions over a certain amount should trigger a warning.

But these rules lead to a lot of false alarms while actually criminal transactions get lost in this noise. More recently, smaller, newer companies in particular have been trying to use machine learning approaches to identify normal financial activity patterns and only issue a warning when outliers are detected. These are then rated by people who reject or approve the alarm.

With this feedback, the AI ​​model can be optimized so that it adapts over time. Featurespace, a company based in the USA and Great Britain, and Napier, which is also based in the UK, are developing hybrid approaches with which warnings generated by an AI can be corrected and converted into new rules for the overall model.

The rapid changes in behavior in recent months have highlighted the benefits of more adaptable systems. Financial regulators around the world have issued new guidance on the types of activity anti-money laundering teams should look out for. But it was too late for many, says Araliya Sammé, head of financial crime at Featurespace.

“When something like Covid happens and everyone’s payment patterns suddenly change, you don’t have time to introduce new rules. If you discover something and alert the people who need to know, the money is gone. ”You need technology that can catch the problem as it happens.

According to Dave Burns, Napier’s chief revenue officer, Covid-19 suddenly made a long-smoldering problem catch fire. “This pandemic was the turning point in many ways,” he says. “Some of the bigger players in the industry have been caught off guard.” However, that does not mean that you simply implement the latest technology. “You can’t do AI just for AI’s sake, as it just spits out garbage,” says Burns. Rather, every bank and payment provider needs a tailor-made approach.

Anti-money laundering technologies still have a long way to go. But the pandemic has uncovered cracks in existing systems that are worrying people, Burns says. That means things could change faster than they would otherwise have. “We see a greater urgency,” Burns. “What are traditionally very long, bureaucratic decisions are now being accelerated dramatically.”


To home page

Leave a Reply

Your email address will not be published. Required fields are marked *

error: Content is protected !!