The New York Times reports today
that the Pentagon has decided to sharply curtail the scope of its "Total Information Awareness
" project, limiting it to only data already held by government agencies. Dr. Tony Tether, directed of the Defense Advanced Research Projects Agency (DARPA
), told a Congressional subcommittee that the change was being made to avoid civil liberties problems that had led Congress to legislate limits on the project on February. The legislative limits entailed a reporting requirement which prohibited the Pentagon from moving forward with the project until they developed a plan to mitigate any civil-liberties risks. Dr. Tether's comments indicate that the agency has decided to sidestep these risks by using data -- like crime records -- that are already in government hands.
... the official said the program, the Total Information Awareness program, would rely mostly on information already held by the government, especially by law enforcement and intelligence agencies.Analysis:
* * *
Today, under friendly questioning by Representative Adam H. Putnam, a Florida Republican who is the subcommittee's chairman, Dr. Tether said the main area of private data that might be useful in anticipating terrorist attacks would be transportation records, since terrorists had to travel.
Saying "I'm trying to help you guys a little with your p.r. problem," Mr. Putnam invited Dr. Tether to swear that the agency was not "contemplating" using credit card, library or video-rental information. Dr. Tether said he could see no value in any such data, but he could not swear that no consultant hired by the agency was not "contemplating" the value.
* * *
Dr. Tether argued that from the outset of the Total Information Awareness project, Darpa had been aware of the need to protect privacy. One essential element was concern by different agencies that sources of their information be kept secret. "Historically," he said "agencies have been reluctant to share intelligence data for fear of exposing their sources and methods."
This last part is the most important. TIA is not about spying on American citizens or setting up some massive Big Brother apparatus. It's about bringing together information and using it more efficiently and intelligently. Have you ever seen a police background check on TV? Do you think it's really that easy? In reality, it's exceedingly hard to do a good background check on someone -- even if you have them in custody. No agency has the ability to call up information on demand -- even information like criminal records, which it should already have. Across the 50 states, police agencies and other agencies use varying databases to keep their information. Some of it has been put together, such as the National Crime Information Center
(NCIC) system. But for the most part, it remains separate. TIA is about bringing this data together so that police officers and security agencies can use it more effectively.
There was some initial conceptual talk about gathering other
data -- such as credit information, financial data, travel data, etc -- to make the TIA database even more comprehensive. The point of this was to gather the data points most relevant to international terrorism. Terrorists don't act like criminals, and they rarely have prior records or speeding tickets that can create a pattern of criminal activity. The salient indicators of terrorist activity tend to be financial in nature, or travel-related in nature. Thus, Pentagon officials designed the project to incorporate those data sources. However, this raised a firestorm of controversy, and many objected to the government's collection of this information. (Query why they'd rather let TRW or Equifax manage this data than the Pentagon) Thus, it's been cut out. I think this is a mistake, because this data remains invaluable for the detection and prevention of terrorist activity.
But perhaps we're seeing the first baby step here towards TIA, and the Pentagon needs to "proof" the concept first before the American public will accept the use of this data. The purpose of all DARPA projects, initially, is to prove the concept. But now, the Pentagon also needs to make the TIA project as criticism-proof as possible, if it has any hope of survival. The DARPA folks are testing TIA for their own security-oriented purposes, and that of other interested agencies (like DHS). But they also need to demonstrate this system to the public, in order to build public trust in the system and answer the criticisms thus far that it will infringe on civil liberties. More to follow...Update:
Eugene Volokh has some interesting thoughts
on where this all may lead, based on his work
on the dynamics of slippery slopes in law and policy.
...the slippery slope may well be in operation here -- I'd call it a mix of the simple attitude-altering slippery slope, either an erroneous evaluation slippery slope or an accurate evaluation slippery slope (depending on whether you think the public will correctly estimate the value of the project), and cost-lowering slippery slope (in the sense that providing the government with more experience about how to run TIA-type projects will lower the cost of broadening them in the future). But some slippery slopes are good, if you already think that what's at the bottom of the slope (e.g., the broader TIA) is good, or if you're not sure whether it's good but you think the first step might indeed provide useful information about the likely value of the next step, and you think that the political system will act soundly based on that information. And of course it's possible that the slippage won't occur, because the public will maintain the line between the government simply organizing the data that it already has, and the government also incorporating data from private sources.
I think he's right -- these initial steps for TIA are designed to make the next ones easier. Starting TIA with a small data set is not intended to limit the project's eventual scope. Rather, it's designed to assuage concerns today about TIA, so that people will object less in the future to the project when it does acquire more data. If you object to TIA in its current prototype form, with its present data set, the time to speak up is now. If TIA proves itself with the government-owned data set, it will most likely be expanded to include medical information, credit information, and other indicators which may be valuable in the national security context. The slope becomes both steep and slippery from here.