Homeland Security

DATE=4/20/2004

TYPE=DATELINE

TITLE=DATA-MINING AND TERRORISM

NUMBER=4-0180

BYLINE=BARRY NEWHOUSE

EDITOR=CAROL CASTIEL

PHONE=619-1101

DATELINE=WASHINGTON

Voiced version and Inserts are available in Dalet - SOD/English News Now/Reports

DISK: DATELINE THEME [PLAYED IN STUDIO, FADED UNDER DATELINE HOST VOICE OR PROGRAMMING MATERIAL]

INTRO: Law enforcement officials in the United States are using powerful new computer programs to look through millions of records, searching for information about terrorists. So called "data-mining" software searches for suspicious patterns in information stored in electronic databases. By quickly scouring criminal files or airline passenger lists, proponents say data mining has the potential to revolutionize detective work. But the powerful technology is also controversial. In this edition of Dateline, V-O-A's Barry Newhouse reports on efforts that use data mining in the fight against terrorism.

TEXT: There is little doubt that in the twenty-first century, information is power. Advancements in the ability to access and store information have revolutionized everything from financial markets to the military. Computers have played an essential role in this information revolution by dramatically increasing the ability to store vast amounts of data. With powerful databases and technology like the Internet, never before has so much information been available to so many people. But with the huge increase in accessible data, there is a growing need for tools that can make sense of it all. Interpreting massive amounts of information is what data-mining software does best.

ACT 1 SHEARER: It's used really anywhere that people have data and where they have problems they need to solve [some sort of problems] related to that data. So what it's about is helping people understand what's hidden in data, hence the term "data-mining" - we dig the useful things out of the data.

/// END ACT ///

Colin Shearer led the team that designed the world's most popular data-mining program, named Clementine. By using complicated mathematical formulas and statistical analyses that sift through huge data sets, Clementine assists in a wide variety of applications, from developing marketing strategies to analyzing medical trials and catching criminals.

The police department in Richmond, Virginia has been using Clementine for more than two years to track crime, streamline department operations and help detectives solve cases. Colleen McCue oversees the department's crime analysis unit and says the software is significantly changing law enforcement.

ACT 2 MCCUE: I think it really is a paradigm shift in terms of how we analyze public safety data. It really does take us - gives us the ability to put more science and less fiction into the analysis and it takes us away from counting crime to beginning to predict and anticipate it.

/// END ACT ///

Despite praise for the software's potential, it also has pitfalls. Colin Shearer says, although data-mining can find useful information in massive databases, it shouldn't be relied on too heavily to draw conclusions about individual cases.

ACT 3 SHEARER: So for example if you're looking through a huge database of people's behavior, looking for activities that may indicate some kind of criminality, just because you find someone unusual doesn't mean the systems can say, yes that person is fraudulent. All it's saying is, this is different. So I think if people expect it to go the whole way through and to make decisions for them and say is that this person is a criminal or a potential terrorist, this person isn't. They're going too far.

/// END ACT ///

Critics of law enforcement's use of data-mining fear that relying too much on the software could criminalize innocent people. They worry that as data-mining programs increase in size, so does their potential for error or abuse. Several new federal anti-terrorism programs plan on using data-mining techniques to analyze huge national criminal databases, airline passenger lists and other types of data. Security analyst Bruce Schneier fears the possibility of mistakes.

ACT 4 SCHNEIER: I think when you have billions of people on the planet you're going to have lots of false alarms. You won't be able to find the signal. It's not a question of terrorism - we've been trying to do this in marketing, we've been trying to do this in all sorts of fields. Data-mining works, but it works very sloppily. And if you want to send someone an offer for a discount on a car, that's ok. But if you want to arrest somebody, that's a big problem.

///END ACT///

There is also opposition to anti-terror data-mining programs that could collect private information. Privacy advocates say the government should not have wide access to people's private financial and personal records because of the potential for abuse.

Tom Mitchell is a counter terrorism technology expert at Carnegie Mellon University in Pittsburgh, Pennsylvania.

ACT 5 MITCHELL: I don't mind that there's data about my communications sitting right here on my computer. That doesn't make me nervous because I believe the data is under my control. And so I think a large part of it is I want to know who is getting to see that data about me.

/// END ACT ///

But even critics of the proposed anti-terror projects say they still think the technology is useful. And one government-sponsored data-mining project has been praised by both security analysts and privacy advocates.

Researchers at the University of Maryland are putting together a database of more than 70-thousand terrorist attacks that took place over the last 30 years to better understand the phenomenon of terrorism.

ACT 6 LAFREE: Everybody has an opinion on terrorism, there's all sorts of stuff in the press about terrorism. It's an odd area though because I think you'd be hard pressed to find a field where so much has been written on the basis of so little hard evidence. And so we're really trying to contribute to that debate by providing some evidence that you know is not perfect but I think is at least better than the data that has been out there.

/// END ACT ///

Gary LaFree heads a Justice Department-funded project that is building a vast public archive of news reports and political and economic data. Because the U-S law enforcement databases aimed at catching criminals are classified and focused on finding individuals, Mr. LaFree says researchers can't use them to study larger trends.

ACT 7 LAFREE: Even though you could have detailed information on every terrorist suspect and every terrorist arrested in the world, that's going to tell you nothing about the economic impact of a terrorist event on tourism or G-D-P or unemployment.

///END ACT///

Mr. LaFree says he's looking for a descriptive image of global terrorism--its impact on governments and law enforcement, and how it has evolved and changed over time.

The main difference between databases like Mr. LaFree's and the more controversial ones proposed by the federal government, lies mostly in where they get their information. Mr. Lafree uses only publicly available data obtained mainly from press reports. He says public sources can offer valuable insights into terrorism because, unlike most criminals who want to hide their crimes, terrorists want publicity for their crimes - usually through the news media.

ACT 8 LAFREE: Because terrorists have this major goal of wanting to get into the media, I think we're likely to pick up a fairly high proportion of major terrorist events. So in some ways we're in better shape than say money laundering which people are actively trying to suppress and certainly don't want media attention.

/// END ACT ///

Mr. LaFree and other scholars say because of the serious lack of hard data about terrorism, any available hard data that they can analyze increases their knowledge.

While the debate over data-mining programs used by law enforcement agencies continues, experts say the technology behind the controversial programs is an essential tool in the effort to learn about and prevent terrorism. Even critics say data-mining is here to stay, but say they hope the public stays engaged in the debate over how the powerful technology is used. For Dateline, I'm Barry Newhouse.

VNN/BN/CC