-- Threat to Civil Liberties or Method of Ensuring Accurate Law Enforcement?
Last week, the Senate amended an authorization bill to include a reporting requirement for the Pentagon. The Defense Department may only proceed with its "Total Information Awareness
" project if it submits a report to Congress on the details of the project, and the measures it's taking to protect civil liberties in the project. TIA is the massive database project sponsored by the Pentagon which would bring together information from all sources (private, public, intelligence, criminal, etc) to support national-security and criminal-investigative authorities. On Friday, House Speaker Dennis Hastert echoed the Senate's sentiments, saying he would endorse a similar amendment in the House. Hastert, R-Ill., is concerned about the privacy implications of the research program, called Total Information Awareness, Hastert spokesman Pete Jeffries said. He said it remains unclear who will fight for the project when House and Senate lawmakers meet next month to decide its future. "Its fate is questionable," Jeffries said.
This makes it all but certain that such a reporting requirement will be added. In general, I agree with reporting requirements
and the goals of Congressional oversight -- particularly for the Defense Department. (Some reporting requirements are quite odious though. A cursory read of Title 10, United States Code, will reveal hundreds of extraneous reporting requirements for the Pentagon that devour thousands of manhours to complete -- and are probably never read by Congress.)
However, TIA may be a good thing when it's eventually developed. I think the debate over TIA has been skewed, first by the William Safire NY Times column
which broke the story, and subsequently by civil libertarians who have cast the issue as one of "Big Brother" instead of as a means to more accurately focus law enforcement. Here are the reasons why I think TIA is a good thing, and why the critics are wrong.
1. DARPA (Defense Advanced Research Projects Agency) does "deep" research
, not near-term or short-term research. They develop ideas & concepts that may come to fruition in 30 years. Many of these ideas, in their infancy, are susceptible to misinterpretation because they are so powerful, and because the science they propose is so revolutionary. But we ought not discourage the basic research sponsored by DARPA because of its possible effects a generation in the future. We ought to instead develop control measures for that risk, and ensure the research does not lead to such harm.
Example: Among other things, DARPA created the Internet in the 1960s as a way for four universities (Berkeley, MIT, Stanford and UCLA) to communicate about sensitive defense-related research and technology. The interconnected network was designed to withstand a nuclear attack, because it would reroute itself around any destroyed links. Thirty years ago, no one could have predicted the evolution of DARPA-Net to the Internet of today. But imagine what would have happened if civil libertarians then had criticized it because of its potential for surveillance, or if movie studios had criticized it because they foresaw some Intellectual Property problems. The Internet might never have been built. TIA may have the same potential for the future, or it may have problems we don't know about. Scuttling it or any other DARPA project now, before it's designed or built, may chill research and innovation with the long-term potential to improve our lives.
2. American law enforcement agencies have a real problem with racial profiling
; TIA may solve part of that problem. They often use race as a proxy for sweeping up terrorists because they lack any meaningful information about who is or is not a terrorist. Part of the problem stems from "indicators", as the Intelligence Community calls them. Indicators are pieces of information which mean things, and when analyzed, can indicate the presence or absence of something. Terrorists leave pieces of information which can indicate terrorist activity. But often, these indicators are relatively inocuous by themselves -- taking out large sums of cash, buying airplane tickets, taking self-defense lessons -- these things mean nothing by themselves, and indeed are activities that lots of innocent people do as well. Without knowing which indicators are important, and a system for tying together indicators from all sources, there is simply no way to put together the dots. A better system of gathering information and analyzing information will produce more accurate law enforcement. It will tell law enforcement what things to look for. Instead of looking for things like race, they will look for precise details of behavior that are the most probable indicators of terrorism. In theory, this will lead to greatly reduced need for racial profiling.
3. A related problem is the gathering/collation/cross-referencing of information
. There currently exists no single system in America for gathering criminal-investigative information from local, state and federal agencies -- let alone combining such information with data from the Intelligence Community, State Department, or foreign agencies. This was cited, among other things, as a key failure in the Senate Select Committee on Intelligence's report on the failures of the Intelligence Community to prevent 9/11
. TIA may represent the answer to this problem, by providing a relational database with sufficient breadth, depth and sophistication to bring together data from all these sources. As a footnote, Oracle already makes such databases for the private sector, and they are used quite well by credit bureaus and other private actors. This system would build on that technology, but attempt to make the system even more reliable (we all know that credit reports can have flaws sometimes). Gathering this information is absolutely critical. Terrorists fly below our radar because they show up as ordinary members of society. The clues they leave are usually innocuous, and tend to fall within the jurisdictions of separate agencies. (This is intentional -- al Qaeda doctrine teaches terrorists to exploit seams like federal/state, foreign/domestic, military/civilian, and to cross these lines when expedient.) America must have a system for gathering this information from all sources and putting it together. The amount of data is too large for a human analyst to deal with -- only a sophisticated computer database can put all this information together to find the right indicators of terrorist activity.
4. The system still has checks
. TIA will not produce some Minority Report
-like world where three pre-cogs sit in a vat and make decisions that lead automatically (without trial or evidence) to lifetime incarceration. TIA doesn't even provide enough information for "probable cause", as I understand the technology. Instead, TIA enables law enforcement to focus their scarce resources on where it really counts. Instead of surveilling the entire Arab-American population (a uniformly dumb idea), TIA would enable the FBI to surveil only those persons with certain indicators of terrorist activity (like repeated trips to certain countries or ties to certain individuals on the State Department's watchlist.) Ultimately, the government must still prove its case in court to incarcerate someone, and it must still try them before a jury or judge in most cases. The Article III courts stand as a bulwark against any slippery-slope problem here.
This is a tough issue, but one I think we need to resolve in favor of Total Information Awareness. The technology promises benefits that can barely be imagined right now, with costs that can be controlled and mitigated. In any case, I welcome your thoughts & feedback
on these ideas.