The Wall Street Journal, September 13, 1995, p. B2. Software May Dry Up Money Laundering By Margaret A. Jacobs Can artificial intelligence be used to combat crime by ferreting out money laundering? Officials at law enforcement, defense and intelligence agencies like to think so. They have suggested creating a sophisticated computer program to screen records of the more than 700,000 electronic money transfers involving U.S. institutions each day and to flag suspicious ones for further investigation. By using so-called artificial intelligence, they hope to stop some of the $300 billion in profits from drug deals and other illegal activities that they estimate is laundered through financial institutions world-wide each year. But in a report issued yesterday, the congressional Office of Technology Assessment says any such plan would face considerable obstacles. Federal privacy laws generally prevent banks and other financial institutions from sharing with the government much of the information they have about electronic money transfers, unless the government issues a subpoena or obtains a search warrant, the report notes. The report also cites technological limitations,, including the difficulty of developing a profile of suspicious activity that a computer could search for, and the high cost of such a program, among other things. Though Congress has yet to consider proposals to use artificial intelligence to fight money laundering, the Senate Government Affairs Committee -- at the urging of defense contractor Mitre Corp. of Reston, Va., which has expressed interest in supplying such a system -- had requested that the OTA convene a panel of experts to study the matter. The OTA advises Congress on technology issues, but isn't expected to receive funding in the next fiscal year. Banks and other financial institutions are now required to report to the government only cash deposits exceeding $10,000. The institutions aren't required to report electronic transfers, even though the OTA estimates that the vast majority of money is laundered that way. Such transactions typically involve dozens of transfers made through banks in various countries, often in the names of shell corporations, with the intent not to leave a trail. An estimated one-tenth of one percent of all electronic transfers involve funds being laundered. By next spring, however, U.S. banks will be required by Treasury Department regulations to retain more-detailed records of each electronic transfer, including the identity of the sender, and to report to the government suspicious electronic transfers that come to their attention. Artificial intelligence is "clearly where we have to be headed," says Stanley E. Morris, director of the Financial Crimes Enforcement Network, a Treasury Department agency that oversees anti-money-laundering efforts. But the OTA notes that giving the government access to the records of individuals is impossible under current law, unless the government issues a subpoena or obtains a search warrant, much like it must do to tap telephones. To obtain a search warrant and to enforce a contested subpoena in court, the government has to convince a judge or magistrate that there is a likelihood of criminal activity. While the law doesn't give corporations as much protection, most banks don't turn over corporate money-transfer records without a subpoena or search warrant. The OTA also notes that banks fear that allowing law enforcement unrestricted access to electronic-transfer data would make businesses in Europe and other industrialized nations reluctant to use U.S. financial institutions. The U.S. would "run the risk of the flight of legitimate capital" from domestic banks, the OTA concludes. "For any artificial-intelligence system to work we also would need the same level of surveillance," carried by other countries, says Joel Reidenberg, a law professor at Fordham Law School and a privacy expert who advised OTA on the report. "Most other countries generally have more, not less privacy protections than we do," he adds, noting that such protections are being strengthened in Europe. Businesses are also concerned that if the government requires financial institutions to retain more data on electronic money transfers, the information will be exploited by banks or revealed through litigation, possibly giving competitors a look at their corporate planning, stock strategies and vendor relationships. "Once banks have this vast amount of information, there will be tremendous pressure on them to use it," says Prof. Reidenberg. A system without strict safeguards against unauthorized disclosure would create a "surveillance state that a democratic society just won't accept," he adds. But even if privacy concerns could be overcome, the OTA and electronic moneytransfer experts say it isn't likely that the technology could deliver. One big problem is deciding what types of transactions are suspicious -- a key step in developing a computer-screening system, says John Byrne, senior federal counsel with the American Bankers Association. For example, the OTA notes that a computer might look for five $3,000 transfers in a single day on the assumption that the intent was to evade cash-reporting requirements. But without a more sophisticated profile, such a search would target many legitimate transfers for investigation, broadly intruding into the privacy of innocent companies and individuals, according to the OTA and privacy experts. The OTA recommends that the government more aggressively use subpoenas and search warrants to expand its knowledge of the way money launderers use electronic transfers before embarking on anything more than a limited artificial- intelligence project, among other things. Meanwhile, some big banks are understood to be experimenting with artificial intelligence on their own. Systems used internally at banks aren't likely to raise privacy concerns and may help fulfill federal requirements to report suspicious activity to the authorities, as well as reduce losses related to money laundering. Last year, American Express Bank International, an American Express Co. subsidiary, agreed to pay a $7 million fine as part of a settlement with the Justice Department of a civil money-laundering case. The department contended that the bank failed to cooperate in the criminal prosecution of two former employees who were convicted of aiding a Mexican drug dealer in laundering about $30 million. [End] -- +---------------------+--------------------------------------------------+ | ____ ___ | Justin Lister ruf@cs.uow.edu.au | | | \\ /\ __\ | Center for Computer Security Research | | | |) / \_/ / |_ | Dept. Computer Science voice: 61-42-214-327 | | | _ \\ /| _/ | University of Wollongong fax: 61-42-214-329 | | |_/ \/ \_/ |_| (tm) | Computer Security a utopian dream... | | | Disclaimer: dreaming is at own risk | +---------------------+--------------------------------------------------+