United States Department of Defense
DoD News Briefing[Excerpt on DARPA'S Total Information Awareness program]
November 20, 2002
(Presenter: ASD(PA) Victoria Clarke. Also participating was Navy Rear Adm. David A. Gove, deputy director for global operations, J-3, Joint Staff and Under Secretary of Defense for Acquisition, Logistics, and Technology Edward C. "Pete" Aldridge.)
Q: Can you help us to better understand what Admiral Poindexter's operation is all about, and how far along he is in developing his program or plan or -- (inaudible)?
Clarke: I can't, but I have someone here who can. (Laughter.) And Undersecretary Pete Aldridge, who was thinking: Okay, I've been here for a while, time for me to leave -- (laughter) -- would be happy to address that question.
Q: What's a nice guy like him doing in place this?
Clarke: That's right.
Aldridge: I asked the same question.
Well, I -- we anticipated that this issue may come up, so I have prepared a very short statement, and then if that statement doesn't clarify what we're trying to do, I'll stay up here for a few minutes for some questions.
My statement goes along the following: The war on terror and the tracking of potential terrorists and terrorist acts require that we search for clues of such activities in a mass of data. It's kind of a signal-to-noise ratio. What are they doing in all these things that are going on around the world? And we decided that new capabilities and new technologies are required to accomplish that task. Therefore, we established a project within DARPA, the Defense Advanced Research Project Agency, that would develop an experimental prototype -- underline, experimental prototype, which we call the Total Information Awareness System. The purpose of TIA would be to determine the feasibility of searching vast quantities of data to determine links and patterns indicative of terrorist activities.
There are three parts to the TIA project to aid in this anti- terrorist effort. The first part is technologies that would permit rapid language translation, such as you -- as we have used on the computers now, we can -- there's voice recognition capabilities that exist on existing computers.
The second part was discovery of connections between transactions -- such as passports; visas; work permits; driver's license; credit card; airline tickets; rental cars; gun purchases; chemical purchases -- and events -- such as arrest or suspicious activities and so forth. So again, it try to discover the connections between these things called transactions.
And the third part was a collaborative reasoning-and-decision- making tools to allow interagency communications and analysis. In other words, what kind of decision tools would permit the analysts to work together in an interagency community?
The experiment will be demonstrated using test data fabricated to resemble real-life events. We'll not use detailed information that is real. In order to preserve the sanctity of individual privacy, we're designing this system to ensure complete anonymity of uninvolved citizens, thus focusing the efforts of law enforcement officials on terrorist investigations. The information gathered would then be subject to the same legal projections (sic) currently in place for the other law enforcement activities.
Aldridge: Protection. Legal protections.
It is absurd to think that DARPA is somehow trying to become another police agency. DARPA's purpose is to demonstrate the feasibility of this technology. If it proves useful, TAI [sic: TIA] will then be turned over to the intelligence, counterintelligence and law enforcement communities as a tool to help them in their battle against domestic terrorism.
The bottom line is, this is an important research project to determine the feasibility of using certain transactions and events to discover and respond to terrorists before they act. We all share the frustration associated with vague warnings of terrorist threats. We hope that TIA will help the U.S. government narrow those generic -- genetic reports -- generic reports down to advance notice of specific threatening acts. I hope that's clear.
Q: There are two things that bother a lot of people -- one, the "Big Brother" aspect, and if you can talk about possible checks and balances -- the second thing is the choice of the man to lead it. I mean, Admiral Poindexter was under a cloud. You know, he was a convicted felon, even though the conviction was overturned on appeal, for lying to the Congress. Is he the kind of guy you'd really want in a situation like this, who has a record of lying and handling untruths?
Aldridge: I'll repeat, again, that what John Poindexter is doing is developing a tool. He's not exercising the tool. He will not exercise the tool. That tool will be exercised by the intelligence, counterintelligence and law enforcement agencies. So --
Q: Why choose him? There are lots of people available who could have run that organization.
Aldridge: John Poindexter has a passion for this project. He's a Ph.D. in physics. He has an enthusiasm for the project. He came to us with the project after September the 11th and volunteered it to DARPA. That was briefed -- Tony Tether, the director of DARPA, came over with John and briefed it to me, and I thought it was a project worthy of the pursuit of the Defense Advanced Research Project Agency. And you want an enthusiastic leader. Once the tool is developed and -- John will not be involved. But it's his enthusiasm and his volunteering of this idea which is why we developed and started to fund it.
Q: What about the checks and balances?
Aldridge: The checks and balances will be the normal checks and balances through the law enforcement agencies who will be exercising the tool, as they do today. There's no difference.
Q: (Off mike) -- tool to the intelligence community, though. A large part of the intelligence community is connected to the Defense Department. Isn't that -- won't they then be involved in domestic law enforcement to some degree?
Aldridge: No, they'll be involved just as they are today in enforcing privacy laws. There's not going to be any difference. This is just another tool to allow them to exercise their ability to go track where terrorists are and to prevent terrorist acts with certain kinds of technology. So that's the problem.
Q: I'm sorry, I don't understand one piece of this. Is this entire program based on totally fabricated data? In other words, it's all hypothetical?
Aldridge: There's some real data that we use, but it's normal data that's available legally. The privacy issues, those will be fabricated stuff.
Q: Okay. That's what I didn't understand. Can you help us understand what is the fabricated data, what is the real data, and what are the privacy issues if you're using fabricated data?
Aldridge: There are no privacy issues. We will not use any data that affects -- that will have any relationship at all to privacy issues. Most of the data is synthetic. It's generated just to exercise the analysis. The real data will have to come from the agencies themselves once they have the tool.
Q: Can you give us some examples --
Aldridge: There will be no privacy issues. This will be all data that will be available in the open and fabricated, to use. There is nothing dealing with individuals at all in this particular exercise, in this feasibility study. It's all generated data for the purpose of the exercise.
Q: So what kind of real data are you using that you just mentioned?
Aldridge: I will have to find -- I don't know the answer to that question, exactly what kind of data. I'm not into the details of the thing. But I don't think there is a problem with it at all.
Q: Can you run over the transactions again? It sounds like every time I would enter or a citizen would enter a credit card, any banking transaction, any medical -- I go see my doctor, any prescription, all of those things become part of this database -- right? -- hypothetically?
Aldridge: Hypothetically they would, although the data that would go along with personal information such as bank accounts, that would all be protected in the Privacy Act just as it is today. Individuals would not be associated with that.
Q: So you need rapid language translation because you are trying to tap into databases of other nations, if they will allow that? Is that --
Aldridge: Or -- yes. Exactly.
Q: And does any of this involve collaborating or connecting, for example, takes from signals intelligence into the rest of this database?
Aldridge: I'm not going to get into the use of intelligence data. But you can be assured that the databases we're trying to investigate -- again, as the feasibility, will this all work, and try to take as much information as we can. When a person enters the country, for example, a visa that comes into the country, you'd like to have that in the database. If they apply for a gun license, you'd like to have that in the database. If they buy a certain amount of chemicals or apply for a gun permit, I guess --
Q: Every time they use a telephone, that call enters the database. And if it is voice recognition, for example, then that enters the database, hypothetically, right?
Aldridge: Hypothetically, yes.
Q: How is this not domestic spying? I don't understand this. You have these vast databases that you're looking for patterns in. Ordinary Americans, who aren't of Middle East origin, are just typical, ordinary Americans, their transactions are going to be perused.
Aldridge: Okay, first of all --
Q: And do you require search warrants? I mean, how does this work?
Aldridge: First of all, we are developing the technology of a system that could be used by the law enforcement officials, if they choose to do so. It is a technology that we're developing. We are not using this for this purpose. It is technology.
Once that technology is transported over to the law enforcement agency, they will use the same process they do today; they protect the individual's identity. We'll have to operate under the same legal conditions as we do today that protects individuals' privacy when this is operated by the law enforcement agency.
Q: So they would need a search warrant, then?
Aldridge: They would have to go through whatever legal proceedings they would go through today to protect the individuals' rights, yes.
Q: As part of this feasibility study, will anybody be looking at legislation, regulation, executive orders that may need to be modified?
Aldridge: I think that's probably an issue that's going to be taken up by the new office of homeland security, who probably will be very much involved in this type -- the use of this type of information.
Q: What is the time line for completion of this technology? And once it is completed, will you give the public sort of a more elaborate explanation of how it works, and maybe even a demonstration of this sort of technology to deal with some of these concerns?
Aldridge: In fact, I was reading somewhere the other day, I think Senator Gary Hart said this is a project of $200 million a year. I don't know where he got the number. The project is funded in the fiscal '03 budget by the president at $10 million. We're in the process of developing the '04 and out-year budget as we're in the process right now. We don't know where that's headed at this point. And the feasibility, it's several years away, based upon the ability to understand the technology.
Q: Several years away before this tool will be available, in other words?
Aldridge: Yes. Yes.
Q: Can we stay on this subject before you get off that, please?
Q: You described one of the functions as to establish connections between transactions. Well, that sounds --
Aldridge: And agencies.
Q: Right. Well, that sounds like a perpetual fishing expedition, as opposed to something for which a search warrant would be sought. For example, if subject A withdrew a lot of money and bought a crop duster, and then over here, bought chemicals that aren't normally used for crop dusting, that's what sounds like you're after. And you wouldn't necessarily have a specific search warrant for that kind of information.
Aldridge: I think that's a good point. Because what we're really looking for -- if you were a terrorist, and you wanted to conduct a terrorist act, you would undertake certain kind of actions, transactions to do that. One, you have to enter the country, and you would probably buy -- get a driver's license, or you would maybe take lessons in airplanes, or something like that.
You're looking for trends in transactions that are associated with some potential terrorist act; that's what you're looking for. And you're trying to put those pieces together. And I -- what this is trying to do, is can this technology work to the point where we as Americans could feel a little more comfortable that our country was protected against potential terrorist acts? That's what we're trying to accomplish. The ultimate goal is that, and to prove the technology works.
Q: But if --
Aldridge: I'm going to answer one more, then I'm going to go back to F-22, because I want to answer that question.
Q: -- Again, to follow up on that question. It sounds like the only way it will work is by casting the widest possible net to encompass the broadest possible number of transactions, as opposed to focusing on individuals for whom law enforcement had some specific interests.
Aldridge: I don't know what the scope of this is going to be, what it's going to take to make this work yet. That's what John's trying to find out. Are there things that we can -- do we have to have a huge amount of data to make this work? Or can we work it by looking at the transactions that lead to a terrorist act; they need some understanding of that; and sharing of various pieces of information among all the agencies that deal in this process, so --
Q: Can I just follow up with one thing? Why --
Aldridge: No, let me talk to Tony first. He's been bugging me for an answer to his question.
Q: Can you make it clear, though, and this seems directed more toward foreign nationals coming into the United States and the visa passports that U.S. citizens --
Aldridge: No. No, it's actions, it's transactions that lead to potential terrorist acts; that's what we're trying to get to.
Q: So it could be like a McVeigh renting a truck --
Aldridge: Could be buying a lot of chemicals; if there's somebody buying a lot of chemicals, it looks unusual; buying a gun; all kinds of potential activities that fall --
Q: Buying a gun? Could you flush that out --
Aldridge: I'm just using examples of things that would go along with -- that would be patterns of an individual potentially conducting a terrorist act.
Q: Why is this appropriate research for the U.S. military to be doing? None of the things that you have described here fall under the rubric of -- or under the scope of what the Pentagon does, understanding that you might eventually turn it over to domestic law enforcement. Still, the question is why is this even an appropriate program for the U.S. -- research program for the U.S. military to be involved in? Why not turn it over to the National Institute of Justice or some research element of the domestic law enforcement community?
Aldridge: Well, I think it is appropriate for the Department of Defense. We are in a war on terrorism. We're trying to prevent terrorist acts against our country. We're trying to give our people who understand and try to track down the terrorists with a sufficient set of tools. DARPA, which is a research agency, which has this as a characteristic of trying far-out solutions, has the technical capability to make this work. And I think this is a service to our nation as the Department of Defense has a role in serving our nation.