High-tech tools even more sophisticated than this are now being used as weapons in the war against terrorist networks. As the United States and its allies work to ferret out Al Qaeda’s sources of funding, financial institutions all over the world are under increasing pressure to pitch in. Banks have been required to report transactions that might point to bribery, insider dealing, money laundering and other crimes for years. New legislation passed since September 11 may force brokers, insurance companies and other financial institutions to do the same. But how to sift through tens of millions of transactions each day? With fancy software that costs several millions of dollars a pop. Whether or not it will help put terrorists behind bars is an open question.
The battle against financial crime has historically centered on organized crime and the drug trade, in particular the laundering of drug money. It’s a cat-and-mouse game. In the 1970s U.S. officials required financial institutions to report currency transactions of more than $10,000. When criminals learned to structure their transactions into smaller amounts, the U.S. authorities lowered the threshold. A few years later, they tried to institute tougher “know your customer” rules but were shouted down by consumer privacy groups and financial-industry lobbyists. Even so, banks got the message: they must do more to stop illegal transactions. The mandate became more urgent in 1999 when Russian banks were caught laundering money through the Bank of New York. “It was clear some kind of technological solution was needed,” says criminologist Michael Levi at Cardiff University in Wales.
Some banks turned to software that could look for certain types of account movements known to be associated with criminal activity. Such “rules based” systems help banks trawl through seas of data, but they are merely reactive. “Criminals can quickly figure out what institutions are looking for,” says Peter C. Stockman, an partner at Accenture. More recently, the better-heeled banks have turned to “adaptive” programs, which incorporate various flavors of artificial-intelligence technology developed in the past decade. Neural networks, a programming technique that mimics the observed behavior of neural circuits, give computers the ability to “learn” to recognize new patterns and make new associations from experience. Link analysis is a mathematical technique for assessing interrelationships among reams of data. These and other techniques have made computers think more like detectives–that is, on their feet. They might, for instance, monitor a business account in Germany interacting with another in Sudan. The two accounts don’t always exchange money on the same days, or in the same amounts (which might confuse a rules-based program). An adaptive program might grow suspicious of cash transfers made on a Sunday. And if the money were being consolidated into a third account that had received similar cash transfers, the program might alert a bank compliance officer. Such programs are currently available from the British firm Searchspace, NetEconomy in the Netherlands and Mantas in the United States. Searchspace’s software manages 100 million accounts for the Royal Bank of Scotland, Barclays and other banks. It once flagged an account that had a lot of small transactions of about £15 each, which turned out to be illicit video sales.
A key advantage of adaptive software is that it can turn up the heat on certain kinds of transactions and lower it on others, reducing false leads. This is essential in tracing Al Qaeda’s finances, which are thought to involve small transactions among many small organizations from dozens of countries. A bank might give more scrutiny to money coming to and from countries on the Financial Action Task Force’s blacklist, such as Nauru, the Philippines or Russia, or to correspondent banks, or to suspect organizations such as Islamic charities. If necessary, the software can even profile individuals. “It’s quite possible that banks could up the scrutiny factor on, say, students from Yemen,” notes consultant Martin Feuer at Deloitte & Touche. The technology seems certain to trigger debates over privacy and civil liberties.
Does it work? It’s too early to tell. The Dutch bank ING has used NetEconomy’s software for three years and claims success, but won’t give statistics. The track record of law-enforcement officials in the fight against financial crime is not encouraging. Still, technology “can create a better environment for catching criminals,” says Barry Rider, director of London’s Institute for Advanced Legal Studies. As bankers and cops get better at exploiting adaptive software, terrorists may have fewer places to hide.
