In December, New York Times reporters James Risen and Eric Lichtblau commandeered the news cycle with the revelation that President Bush had approved a program of warrantless wiretaps of domestic-to-international communications by the super-secretive National Security Agency. Days later, they disclosed that the NSA's electronic eavesdropping may have been far more extensive than initial reports suggested, involving the harvesting of enormous volumes of telcommunications data for analysis.
One of their sources was former NSA insider Russell Tice, who earlier this week told ABC News that the communications of millions of Americans might have been vacuumed up in one of a number of classified NSA programs. "If you picked the word 'jihad' out of a conversation," Tice explained, "the technology exists that you focus in on that conversation, and you pull it out of the system for processing."
Critics say Tice is a paranoid, embittered by his May 2005 firing from the agency. Tice says he's a whistleblower who was punished for speaking up—and now wants to make sure Congress hears what he has to say about what he believes are unconstitutional activities carried out in the name of the War on Terror. The programs that have been made public so far, according to Tice, are just the tip of the iceberg. Assistant Editor Julian Sanchez spoke with Tice—and anyone else who may have been listening—by telephone this week.
REASON: You've described technologies capable of sifting through vast numbers of communications and pinpointing very specific information that intelligence analysts are looking for. What can you say about how that kind of technology is being used?
Tice: I can't say how an intelligence agency uses it, because that would be classified. Then the FBI would have shackles and cuffs waiting on me real soon, so I have to be careful what I say. But we can talk about the technologies and we can use hypotheticals and we can use wiggle words.
If you wanted to, you could suck in an awful lot of information. The biggest constraint you're going to have is the computing power you need to do it. You need to have some huge computers to crunch that kind of stuff. More than likely you're talking about picking it up in a digital format and analyzing it depending on how the program is written depending on whether it's audio or digital recognition you're talking about, the computing power is phenomenal for that sort of thing. Especially if you're talking about mass volumes, if you're talking about hundreds of thousands of, say, telephone communications or something like that, calls of people just like you and me, like we're talking now.
Then you have things like, and this is where language specialists come in, linguists who specialize in things like accents and inflections and speech patterns and all those things that come into play. Or looking for key phrases or combinations of key words within a block of speech. It becomes, when you add in all the variables, astronomical.
REASON: Do you have a sense of the scale that's possible, how many phrases and conversations it might be possible to filter?
Tice: Technically it's limitless. It's like, you know what a Boolean logic line is? [Yes.] Think of a Boolean logic line with these sorts of parameters in your normal Boolean, built on these filtering parameters. As long as the software is designed to handle however long the Boolean string is in this case, then you have the computing power and the other equipment to crunch the information to put it through the filtering process. Technically you can do as much as you want. It's going to cost you a lot of money and you're going to have to buy some big computers and other equipment, bit synchronizers and that sort of thing, monitoring error rates.
You have to be careful to overdo it, because if you overdo the situation, you'll saturate your bit error rate. So in our hypothetical situation, you could write a program to do this, but you wouldn't be able to filter enough, say. Ultimately you would have to tweak it over time; you would analyze what your output was and say "no, we're getting too much garbage, so we need to focus on this particular filter or this particular item, to be able to winnow it down to where you want it to be."
You run the risk the other way of omitting information you may have wanted, which is where you need specialists, who know exactly the information you want, to work with the software engineers and the language specialists to make sure that everyone's working in sync so that you get the what you want. Normally a linguist or a software engineer isn't the intelligence analyst or intelligence specialist who knows the nitty-gritty of the intelligence or the information you're looking for.
REASON: There's always a problem looking for low-frequency events in a large population, even with a very good filter. How big a problem do you think false positives are?
Tice: It's going to be a huge problem. Huge. That's going to be your number one concern insofar as false positives are ultimately your error rate. The ultimate goal, more than likely in our hypothetical scenario, is to filter this thing down enough so that you can put it into human analysts' hands. The ultimate filter, the ultimate computer, is the human brain.
REASON: How can you minimize that problem in a system like this?
Tice: Some sort of a built in quality control, and automate that as much as you possibly can. So you're always analyzing the output of this and tweaking it as much as you can. The key in this sort of processing is to get the machines and the computers to do as much of it for you as you can as effectively as you can. Garbage in, garbage out. So all your players have to be working in sync to get something like this together.