AI Fights Money Laundering

group were representatives from the FBI, Interpol, the Financial Action Task Force, and a host of other international and national financial regulators and investigators. 9/11 represented a massive failure of international controls, and clearly only radical change would prevent something similar from happening in the future. What troubled the committee and international agencies was that the whole terrorist operation could have been funded for less than US$0.5 million—pocket change, in the context of routine banking. Up to this point, international money laundering had been about the Mafia, drug smuggling, and arms deals, involving sums in excess of $500 billion a year. 1 How could banks transacting hundreds of millions of dollars a day spot suspicious activity in transaction amounts as small as $2,000 to $5,000? How could they have detected the terrorist financing behind 9/11? Two years later, almost half of the world's top 20 banks are using AI systems, 2 and AI has emerged as the leading method in the fight against money laundering (see the " AI and Money Laundering Detection " sidebar). 3 At Search-space, we monitor customer activity to identify unusual behavior and detect potential money-laundering situations. In 1998, I met with the head of risk for a mid-sized UK bank. Even at that time, the UK had stringent money-laundering regulations owing to the ongoing terrorist threat. Banks would occasionally fall foul of the regulator and need to demonstrate wholehearted commitment to finding ways of trapping suspicious activity. This particular bank was concerned with its staff's ability to successfully monitor such activity, given modern banking's increasingly faceless and electronic nature. Could AI help monitor transaction behavior to detect money laundering? The bank had approached Searchspace, formed by researchers from the Intelligent Systems Lab at University College London in 1993. It applies adaptive and learning-systems approaches to a range of business and finance tasks. However, until then, we had principally developed systems for US and UK stock exchanges to automate market abuse detection—monitoring insider trading, front running, market manipulation, and other forms of market cheating. The problem the bank posed was far more challenging— five million transactions per day, five million individual accounts , over three million customers, hundreds of product types, and no clear signature or pattern associated with money laundering. Unlike many types of financial fraud, money laundering could range from a single transaction to the culmination of months of complex transactional activity. A sequence …