What Are the Implications of Data Mining? A report that the government secretly collected Americans' phone records looms over Gen. Michael Hayden's nomination as CIA director. What kind of information can be gleaned from such records? Is the program legal? Debbie Elliott speaks with reporter Joan Biskupic of USA Today and Artur Dubrawski of Carnegie Mellon University, an expert on data mining.

What Are the Implications of Data Mining?

What Are the Implications of Data Mining?

  • Download
  • <iframe src="https://www.npr.org/player/embed/5403317/5403318" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

A report that the government secretly collected Americans' phone records looms over Gen. Michael Hayden's nomination as CIA director. What kind of information can be gleaned from such records? Is the program legal? Debbie Elliott speaks with reporter Joan Biskupic of USA Today and Artur Dubrawski of Carnegie Mellon University, an expert on data mining.


From NPR News, this is ALL THINGS CONSIDERED. I'm Debbie Elliott. In his weekly radio address today, President Bush urged the Senate to move quickly to confirm Michael Hayden as the new director of the Central Intelligence Agency. That nomination has been under a darkening cloud after Thursday's revelation that the National Security Agency, which General Hayden headed, has been collecting the phone records of millions of Americans. Yesterday, two lawyers filed a class action lawsuit against Verizon, one of the companies that reportedly provided the phone records. Congressional leaders of both parties have called for hearings and said they might subpoena the heads of the phone companies that cooperated with the government. We turn now to Joan Biskupic of USA Today, the newspaper that broke the story of the phone data collection. She's the paper's Supreme Court Reporter. Welcome to the program.

Ms. JOAN BISKUPIC (Reporter, USA Today): Thank you.

ELLIOTT: Let's talk first about the legal issues raised by this program, as reported. Can you just sort of lay out he landscape for us?

Ms. BISKUPIC: Sure Debbie. There's three areas that I'm sure members of Congress will pursue. First of all, you have basic fourth amendment requirements that say people are protected against unreasonable searches and seizures. So the question is, was this an unreasonable search? The second question has to do with federal statues that protect against warrant-less gathering of electronic communications. And finally, the issue is, did the phone companies breach any kind of telecommunications law that says that their customers' information should be protected? So those are the three main areas that are at issue here.

ELLIOTT: Let's start first with this idea that we have this constitutional protection against search and seizure. Is this an illegal invasion of that?

Ms. BISKUPIC: Well, it looks like, as a threshold matter, it would not be unconstitutional under the Fourth Amendment, because what the Supreme Court has said is that there are different kinds of electronic communications here. There's something in the content of our conversations, and that is more protected, what we say to each other. But there's a lower level of protection for pure phone numbers, who we call, the traffic back and forth. And in the late 1970s, the Supreme Court actually ruled that the Fourth Amendment doesn't protect just the pure phone numbers, but then there's another legal hurdle that has to do with what's in federal statutes. The most basic statute at issue here could be one that covers special national security issues. That is the Foreign Intelligence Surveillance Act that essentially says that when the government wants to collect information from our electronic communications that might have to do with matters of foreign intelligence, government agents have to go to a special court and get permission.

ELLIOTT: So we have three telephone companies that have allegedly turned over tens of millions of records and we assume that the government did not go through this process.

Ms. BISKUPIC: Exactly. The reason we can assume that the government did not seek a warrant is because one of the major telecommunications companies, Qwest, that said no to this, had said why don't government lawyers actually go before this special court related to foreign intelligence and say, can we have a warrant to do this? So there is a presumption that the government did not seek a warrant here.

ELLIOTT: Was there anything of this scope that you could find as you dug through presidents?

Ms. BISKUPIC: No. As a matter of fact, the laws that are at issue here, the Federal Intelligence Surveillance Act and laws governing the devices that are used to tap into phones and to trace phone numbers were really written with the idea that government would be going after individuals or government would be going after discrete criminal networks, not massive data collections, as we apparently have here.

ELLIOTT: Now, since it's collecting the phone numbers and we have reason to believe that names are not attached to these phone numbers, so does that change the legalities?

Ms. BISKUPIC: It does in some ways, under the Foreign Intelligence Surveillance Act. It is aimed at content generally, but the definition of communications and identity is fairly broad, so perhaps this so-called FISA court might think, would it be appropriate to pass on numbers in this day and age that can so easily lead to identification?

ELLIOTT: And I guess in this day and age when you can just put a telephone number into Google and find out who it belongs to, it sort of changes things.

Ms. BISKUPIC: It makes it a little more complicated.

ELLIOTT: Joan Biskupic writes about the Supreme Court and other legal issues for USA Today. Thanks for coming in.

Ms. BISKUPIC: Thanks, Debbie.

ELLIOTT: Hours after USA Today reported on the phone record collection program Thursday, President Bush responded to the concerns.

President GEORGE W. BUSH: We're not mining or trolling through the personal lives of millions of innocent Americans. Our efforts are focused on links to Al-Qaida and their known affiliates.

ELLIOTT: To find out how phone logs might help in the fight against terrorism, we turn to Artur Dubrawski of Carnegie Mellon University. He runs the Auton Center, which studies data mining. I asked him, judging from what has been reported, does the government's collection of phone records sound like data mining?

Mr. ARTUR DUBRAWSKI (Carnegie Mellon University): It does, indeed. Data mining is a process of extracting useful information from data and here we are looking at some data that may contain useful information. What we are looking at here is a massive, very complex, web of connections. And data mining technology provides some tools to find interesting patterns in it.

ELLIOTT: Like what?

Mr. DUBRAWSKI: I'm not involved in this particular project or activity, so I can only speculate here. If this is an intelligence agency looks at this data, they may want to find if there is any statistical evidence that some groups of people are linked together. That's there's a likelihood that these groups are terrorist cells.

ELLIOTT: Well, they could just be families, couldn't they?

Mr. DUBRAWSKI: Yeah, oh yeah, I bet that the vast major of this data is about completely innocent people. The agency may have knowledge about some typical patterns of connectivity between terrorist cells in general, and so they may actually try to find whether anywhere those known predefined patterns can be found.

ELLIOTT: Now, this sounds like searching for a needle in a haystack, when you have millions of phone records and you only maybe know the phone numbers of a few suspects.

Mr. DUBRAWSKI: Right, and this actually poses a tremendous technical challenge. And the actual content of useful information in this particular data may be limited. And for practical reasons, I doubt that just phone records could lead us to any solid conclusion. In fact, there are so many other ways of communication available to people these days that if you want to have a really solid picture of connectivity between bad guys, they may want to look at some other sources. Having said that, this data is readily available. It's cheap to get. Phone companies are actually collecting our (unintelligible) information all the time. So from a technical perspective, this is a very attractive opportunity.

ELLIOTT: Artur Dubrawski directs the Auton Lab at Carnegie Mellon University in Pittsburgh. Thank you so much for speaking with us.

Mr. DUBRAWSKI: Thank you very much.

Copyright © 2006 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.

NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.