Exposé  

Totalitarian Information Awareness

by Joe W. (Chip) Pitts III

The nightmare world of George Orwell isn't just fiction anymore, thanks to a new program from George Bush and an old Iran-Contra convict.

---

Unbeknownst to most Americans, their liberty is being consumed by the triad of counter-terrorism programs enacted since Sept. 11: the Patriot Act (considered in these pages in May 2002), last fall's Homeland Security Act, and finally the "Total Information Awareness" initiative conceived and run by former Iran-Contra felon John Poindexter.^*

Total Information Awareness (TIA) starts from the recognition that our daily lives are monitored more than ever before. But while most of us find this at least vaguely objectionable, TIA sees in this an opportunity to capture terrorists. After all, those who visit Times Square in New York, or Piccadilly Circus in London, are on camera hundreds of times per visit; even in less concentrated environments, it's now routine that shopping malls, convenience stores, banks, ATM machines, airports, restaurants and other merchants, and their parking lots capture us as we carry out the transactions that constitute the business of life. Our phone calls are made on mobile networks that include location-monitoring technologies pinpointing where we are at any given time. Palm Pilots, other PDAs, or laptops may inform us of the nearest Starbucks location, but also tell the network where we are and what we want, from the news we read to the books we order online. Satellites hover overhead, communicating constantly with GPS sensors in our cars, trucks, boats, and handheld devices, viewing details as minute as the license plates on our vehicles, as highway toll tags authorize our passage and record our comings and goings. Cable television companies monitor and record our pay-per-view transactions. Travel agencies and our own Internet bookings track our preferences and meanderings. Doctors, hospitals, universities, and insurance companies increasingly rely on networked, digitized medical records. Credit and debit card terminals at grocery and video stores, gas stations, and hotels note where and when we make our purchases, and the digital signals sent along the financial networks determine whether we'll be allowed to proceed with the purchases, or execute our stock trades, or not. The IRS and other government agencies both monitor increasing amounts of such electronic records, and create new records about us and with us, as we pay our taxes, appear in court, or apply for driver's licenses, passports, visas, or other government benefits using electronic means.

TIA would gather up and link all these datapoints and streams — whether emails, phone calls, instant messages, video or audio broadcasts, records of gun purchases, car rentals, or charitable contributions — into hundreds of millions of virtual dossiers on each of us, then "mine" that data for patterns resembling the patterns of terrorist activity coming from other datastreams. In this way, TIA hopes to detect and "preempt" terrorist activity.

If you're not familiar with "data mining," think for example of Internet search engine Google — but on steroids — or perhaps the way those credit card companies sometimes call you because they think they've detected a pattern of fraud in your transaction data (and recall how often they're wrong!). Well, TIA will be the same sort of effort, but orders of magnitude broader.

While the many depredations of liberty over the past year and a half often have been described as "Orwellian," the appellation fits none as aptly as Poindexter's baby. TIA threatens to suck the oxygen from liberty's flame in a way unprecedented in the history of our republic. Even George Orwell, who invented "Big Brother" in his novel "1984," would be shocked at the pervasiveness and ruthlessness of the intrusion that Poindexter is trying to implement.

Poindexter has now removed the logo that graced the TIA website [ed. note: defunct as of 1/25/05] during the first year or so of its existence: the creepy, Masonic, all-seeing eye within a Pyramid, like that on the back of a dollar bill, but emitting a ray of light (or a death beam?) spanning the globe, and accompanied by the slogan "Scientia Est Potentia" ("Knowledge Is Power"). (The prior logo may still be viewed here.) But the ominous connotations of those images linger despite the website's hasty cleansing.

The "Patriot" Act

The infamous "Patriot Act" gutted at least half — the most important half — of the Bill of Rights. The all-important First Amendment's guarantees of free expression and assembly were undercut by allowing an unprecedented degree of government intrusion. Librarians and bookstores are now supposed to turn over information on our reading habits, undoubtedly chilling free thought and breaking down the foundation of free inquiry that undergirds the ability to question and dissent from government actions. The Fourth Amendment's guarantee of freedom in our homes and personal effects from unreasonable searches was seriously undercut by allowing electronic surveillance (without our knowledge) of our home computers, emails, and Internet surfing habits, and "sneak and peak" search warrants authorizing searches of our houses without our being told.

The Fifth Amendment's guarantee of due process and the right to confront witnesses against you means little if immigrants and even American citizens (like Jose Padilla or Yaser Esam Hamdi) can be indefinitely held without charge, access to lawyers, or a judicial hearing, trial, or review of any kind simply by virtue of the government labeling them "terrorists" or "enemy combatants." The Sixth Amendment's right to a speedy public trial apparently applies no longer. Even the Eighth Amendment's guarantee against cruel and unusual punishment seems to be held in little esteem, if certain reports of beatings and solitary confinement of detainees without access to family or health care, and other reports of detainees transferred to face torture in places like Pakistan and Egypt can be believed. While we can be thankful that torture was openly discussed only briefly as a legitimate policy tool, our government's new insensitivity to the delicacies of interrogation methods strongly suggests that it cares far more about ends than means, and about so-called results more than liberties. The Patriot Act sets the stage for intense government focus on the activities of innocent people and misdirects governmental resources from appropriate focus on actual terrorist activities. At least in the Patriot Act some homage was paid to continued weak oversight through the persistence of judicial warrant requirements in some instances (although the process of forum shopping to find a friendly judge renders this more a formal than a truly substantive protection). Since that time, however, a special federal appeals court decision has obliterated most of what was left of the wall between intelligence and law enforcement. TIA's technological approach risks bypassing such protections completely.

The Homeland Security Act

It's ironic, to say the least, that a conservative president putatively committed to limited government has brought into being the most massive new bureaucracy since the various military services were brought into the Department of Defense fifty years ago. The new Homeland Security Act, totaling about 500 pages, consolidates 22 federal agencies and about 170,000 employees, including the Secret Service, the Immigration & Naturalization Service, the Border Patrol, and the Coast Guard. Note that the two agencies most implicated in the intelligence failures preceding Sept. 11, the CIA and FBI, are not included in the agency. Though this is perhaps for very good reason (especially the desirability of separating the different functions of these powerful agencies from the also powerful new mega-agency), it is inexplicable that the CIA and FBI remain essentially unaccountable for their failures.

The biggest reorganization of government in the past fifty years may or may not eventually enhance homeland security, but it certainly doesn't augur well for preserving liberty. Thanks to some strong voices including outgoing Congressman Dick Armey (R-Tex.) — now a consultant with the ACLU, along with his former colleague Bob Barr (R-Ga.) — there are a couple of exceptions to this. The law says that it doesn't authorize either a national identity card, or Attorney General John Ashcroft's previously planned program to turn neighbors, utility repairpersons, the cable guy, the postman, and citizens in general into a "Terrorist Information and Prevention System" (TIPS). This latter was a particularly appalling idea, reminiscent of communist and fascist neighborhood cells reporting on neighbors, reminiscent of Pavlik Morozov, the celebrated 14-year-old Soviet martyr who was honored for being a good party member and turning in his family to authorities.

Alas, the rest of the Homeland Security Act rides roughshod over liberty instead of protecting it. The TIA program contemplates even greater intrusion than the rejected identification card and TIPS programs. The new law creates a "Homeland Security Advanced Research Projects Agency" (HSARPA), modeled after the Defense Advanced Research Projects Agency (DARPA), which created the Internet (sorry, Al). HSARPA is funded at the not inconsiderable taxpayer expense of $500 million annually, and, like DARPA, it will "promote revolutionary changes" in technologies to protect the homeland. What was a separate bill approved by the House of Representatives last summer, the Cyber Security Enhancement Act, was added at the last minute to the Homeland Security Act. It expands the ability of law enforcement authorities to monitor email or telephone conversations without first obtaining a court order. It also grants Internet Service Providers (ISPs), like AOL, Yahoo, or Earthlink, greater room to release information to any governmental authorities if, in their eyes, there's a "good faith belief" that there is an "immediate threat to a national security interest." These liberalizations further loosen the "emergency" exception that had already been provided in the Patriot Act, by removing the requirements that the disclosure be based on a "reasonable belief" and be connected with an "immediate danger of death or physical injury." It places the ISP in the position of distinguishing between evidence of a real threat and (supposedly protected) free speech, with a likely chilling effect on the latter.

Finally, the Homeland Security Act also takes further steps along the road to a more secret government, by expanding the grounds on which TIA or other government activities can be classified, and reducing the scope of access to information about governmental decision-making through the Freedom of Information Act.

The TIA Toolbox

If the Patriot Act sets forth the ideology and plans for a newly intrusive state, and the Homeland Security Act builds the framework, Total Information Awareness may be seen as providing the tools to make it all work.

TIA, moreover, threatens to be just as ineffective an overreaction to legitimate concerns over prior intelligence failures as the Homeland Security Department. But, while the Homeland Security Depart-ment's counterproductive effects are likely to be mainly in the short-run (as culture clashes, lack of system interoperability, and fighting for position diverts officials from vital tasks), TIA is likely to have lasting harmful effects unless modified. TIA remains within DARPA at present, at the vanguard of its attempt to incubate "revolutionary" counter-terrorist technologies. According to DARPA's own documents posted on the Electronic Privacy Information Center website, TIA is funded at $240 million through fiscal year 2003 (despite Defense Department Undersecretary Pete Aldridge's statements on November 20 that the amount was only $10 million per year).

Of course, technology is neutral, with any technology capable of good or bad applications by the user. Some of the voice recognition, automatic translation, and other TIA technologies being developed, like the Internet, have applications that would be socially desirable both in the fight against terrorism and in other private sector uses. For example, the "Babylon" project envisions, on a small and more selective scale, something like the "universal translator" familiar to Star Trek fans: an earpiece combined with a battle-hardened Palm Pilot-like device allowing two-way natural language translation for soldiers in the field, beginning with "low-population, high-terrorist-risk" languages. Given the special difficulty of detecting and quickly responding to and reducing casualties from bioterrorism, the Bio-Surveillance project's work on early detectors, ranging from medical data to "animal sentinels" (operating on the "canary in a coal mine" principle) is also most welcome. The "Genoa" project to facilitate sharing and analysis of data, already legally available, also sounds prudent and helpful, since the failure to predict and stop 9/11 mainly resulted from the government's failure to do just this.

But, as with any technology, placing it in the hands of an untrustworthy or malevolent operator (e.g. Poindexter) raises concerns. A program going by the acronym "EARS" (Effective, Affordable, Reusable Speech-to-Text) aims at more powerful automatic transcription, allowing the government to transcribe more effectively large quantities of audio information — overheard conversations, radio broadcasts, or telephone intercepts — into easily absorbed written text. Another program would automatically extract, summarize, and translate critical information from large quantities of foreign language speech. Still another program, on "Evidence Extraction & Link Discovery," will automatically search Web pages, text messages, phone calls, email, and other data sources, to identify patterns between people that could reveal terrorist activity. "Genoa II" will facilitate "human-machine decision collaboration" to solve complex problems more quickly and reliably. The Human-ID-at-a-Distance program will use cameras and biometric identifiers (e.g. recognizing you by your face, gait, and/or iris, using visible or infrared means) to automatically spot terrorists from afar. (DARPA had also considered but withdrew under pressure a plan to tag all Web communications with biometric identifiers ("eDNA")).

At TIA's core is the "Genysis" project to create the "mother-of-all-databases." The appeal of "anything, anytime, anyplace" knowledge sounds empowering when it's the common vision (as it is) of software, media, and communications companies, relating to individual productivity, control, and effectiveness. The feelings it inspires when inverted to empower government against individuals are not quite the same. God-like omniscience in the hands of a perfect God may be a fine thing, but in the hands of imperfect humans running already-powerful governments, the prospects of such additional power are truly frightening in ways that the American public and lawmakers don't yet appreciate. The long-term harm could be even more serious than the national identification card and the TIPS informant programs outlawed in the Homeland Security Act.

There has never before been an effort to analyze data on the scale contemplated by Poindexter's initiative. "Genysis" aims to develop an "ultra-large, all source, omni-media, virtually centralized, information repository database." DARPA is loathe to suggest it would be anything like the puny commercial "databases" of today's world. (The TIA website says they use the term "database" merely "for lack of a more descriptive term.") The contemplated database would seem to aim at a near infinite capacity, for input of as much data as possible about everyone. A "key metric" described in documents on DARPA's website [ed. note: defunct] is "the amount of total information that is potentially covered." (Great fodder for "Saturday Night Live" and other comedians, but George Orwell might warn you to stifle the laugh.) In case you had any doubt, we're talking the BIGGEST DAMN DATABASE IN THE HISTORY OF THE UNIVERSE. Oracle's Larry Ellison must be filling swimming pools with saliva, drooling over this (now there's a pretty image for you).

This database ("for lack of a more descriptive term") would allow the government to plumb all sorts of public and private information, from all the transactions you engage in, your observed actions, the noises you make, and, yes, even the thoughts you think, to enable the government to rapidly "detect, classify, ID, track, understand, and preempt" terrorists. This would naturally include the manifestations of your thoughts in the email or mail you send, other text you write, books you check out, or websites you surf, but there actually have been and remain government programs to read your thoughts, ranging from the Defense Department's famous psychic weapons detection programs to "non-invasive neuro-electric sensors" being considered for use in aviation security. You may think I'm kidding, but check out this page. They plan to use XML (Extensible Markup Language) or successor languages to facilitate interoperability between all these diverse databases.

Among the disparate databases to be connected (from the TIA website) are financial, educational, travel, medical, veterinary, transportation, housing, government, and communications databases. These will be complemented by "novel methods of populating the database from existing sources" (this already sounds a bit kinky) as well as creation of "innovative new sources." (Blockbuster soft-porn video rental records?) Again, you may think I'm joking, but I'm not: some of the al Qaeda operatives were apparently into porn, and we know that several visited a strip club before 9/11; this database would attempt to extrapolate from the conjunction of such facts with, for example, immigration entry from Saudi Arabia, telephone calls to Afghanistan, and purchases of guns or religious books, to identify terrorists.

I wrote in these pages last year ("A Glimpse Ahead," September 2002) that the administration's "preemption" approaches at home and abroad reminded me of the science fiction writer Philip K. Dick's dystopian visions, rendered on the silver screen in such memorable movies as "Blade Runner," "Total Recall," and most recently, the Tom Cruise-Steven Spielberg collaboration, "Minority Report." One of the TIA programs — "Wargaming the Asymmetric Environment" — really sounds like the Pre-Crime Unit in that last movie: it would aim to "predict and act against" terrorists based on models of past behavior in their political, cultural, or ideological environment. In other words, mere suspicion of possible future terrorist activity based on past activities, including guilt-by-association, could result in present preemptive action. Again, while the specific preemptive methods used aren't tied to the technology and haven't been announced, one can imagine that, in the case of serious terrorist threats, the force used could be extreme. Not that we don't need strong actions against true terrorists: we do. But our Constitution and Bill of Rights are intended to ensure that innocent people are not wrongly harassed and ensnared in the fight against criminals and terrorists. As Marc Rotenberg, Executive Director of the Electronic Privacy Information Center recently told Business Week: "TIA is not likely to be an effective way to prohibit future terrorist acts, but it will have an enormous impact on the government's ability to monitor things not related to terrorism."

Even if TIA's basic projects are feasible (and substantial technical challenges persist, as further discussed below), you're right to ask yourself whether the benefits from such unprecedented government surveillance would outweigh the costs. The only thing scarier than the current TIA plans, which are unlikely to work, is if the government comes up with ways to make TIA work. For, once that occurs, the proliferation of lost liberties will continue, as pressures build for TIA to be turned to other purposes perceived as benign, like fighting crime, collecting taxes, or protecting the rich from the poor. In the meantime, by encouraging excessive faith in technology and overconfident complacency, it's highly likely that this "ultimate system" will result in even more massive intelligence and security failures than those occurring in connection with Sept. 11.

"Trust Us"

TIA is supposed to be merely a research and development project, and in a recent Defense Department briefing, Undersecretary Aldridge stated that the project won't use real-life data. The website and related documents, however, emphasize developing "useable tools" and "leave-behind prototypes" that can be "stress-tested in operationally relevant environments," and both the website and subsequent reports make clear that various agencies actually have and are testing TIA technology in the field, with real data. As the New York Times reported, "a prototype is already in place and has been used in tests by military intelligence organizations."

The program architects say that they will try to "create privacy filters, 'aliasing' methods, and automated data expunging agents to protect the privacy of U.S. citizens, and those not involved with foreign terrorists." The idea would be that data can only be viewed in an "anonymous" fashion unless a certain individual's activity merits closer scrutiny, at which time that individual could be "unmasked." Undersecretary Aldridge says "we're designing this system to ensure complete anonymity of uninvolved citizens." But note the unintended irony in his answer to a reporter's question at his briefing last fall:

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.

Perhaps it was a normal slip of the tongue rather than a Freudian slip, but ordinary people will take little comfort that "the privacy issues . . . will be fabricated stuff," or that proven liar Poindexter is responsible for assuring that such protections are built in to the system. After all, he and Ollie North got into trouble in the first place because they failed in their attempt to delete all copies of 5,000 Iran-Contra emails.

In addition to dispensing comfort that the system design will ensure privacy (though, of course, any system can be hacked by someone, if only the system designers), we're told we can rely on the Ombudsman within the TIA office (who works for Poindexter) or the Chief Privacy Officer or Civil Liberties Officer within the new Homeland Security Department (who works for Tom Ridge in an agency whose mission is security, not privacy or liberty).

Here's Aldridge again:

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 . . . Once that technology is transported over to the law enforcement agency . . . [w]e'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.

So the final bastion that Poindexter and his colleagues point to are the privacy laws in this country — as if the protections offered by those privacy laws were in fact extensive and effective, as if they weren't already substantially eroded in the past year and a half by the Patriot Act then the Homeland Security Act, as if businesses will routinely litigate or stand up to heavy government pressure in the name of security to disclose our information, and as if such laws aren't likely to be even further eviscerated when the next terrorist attack happens (as it surely will). Again, I'm not as much a fan of "Saturday Night Live" as I was in its early days, but if they haven't already done a sketch on such a cruel joke, they should. The sad reality is that U.S. privacy laws have never been very strong, and have recently been weakened considerably. Consequently, some businesses will insist on a government subpoena before complying with government requests for data, but many businesses will roll over in the face of "national security" requests from "the government," and all will be struggling to define the limits of their rights and responsibilities in the current environment.

If the privacy law argument doesn't reassure you, Poindexter, et al. have another one up their sleeve. Invoking the mantle of technology's "neutrality," they assure us that it's not the technology as much as the use of the technology that needs oversight and monitoring. And, they say, effective oversight, e.g. by the agencies using it or by intelligence and armed services committees of Congress, would be more than adequate to protect our lives and liberties. But oversight by one of the agencies using the technology, especially in an administration with as much a penchant for secrecy as this one — which has clamped down on freedom of information act laws and continues to litigate issues such as disclosure of the documents showing Enron's influence on our energy policy — will not be adequate. Neither history nor human nature afford much comfort on this score. One need not invoke Lord Acton's dictum that power corrupts to recall that the history of data records used by defense and intelligence services in this and other countries is replete with instances of records manufactured for innocent purposes, then being warped to nefarious ends by government officials. Hitler routinely used government records in the countries he invaded to determine and eliminate groups and political opponents he didn't like. J. Edgar Hoover, in this country, notoriously abused surveillance and information nominally obtained for national security purposes to persecute communists, suspected communists, anti-war activists, or merely political opponents like Martin Luther King — all despite theoretical oversight from other branches of government. Attorney General John Ashcroft, Presidential spokesman Ari Fleischer, and others in the administration have already suggested that those who disagree with them are aiding and abetting the terrorists.

Undersecretary Aldridge is quick to point out that Poindexter was merely the father and current director of the TIA project, whose "enthusiasm" for the idea "is why we developed and started to fund it." For a hint of Poindexter's enthusiasm, take a look at the logo from his slide presentation describing TIA (see page 25).

Aldridge says that Poindexter would not (necessarily) be the one using these tools. But even if that's true, and who knows whether it will be or not, how many other Poindexters, or Nixons for that matter, might be empowered?

In fact, throughout history, government is responsible for more atrocities and abuses of liberty than any private group. The bottom line is that oversight has never adequately prevented such abuses, especially when there is a real climate of fear insulating government privacy invaders from criticism. As we all know, once technological capabilities exist to ignore or undo safeguards, the temptation to use the technology's full power can be overwhelming. With Poindexter's proven mentality of "the end justifies the means," why would anyone trust him to put adequate protections in place?

While no one would question these technologies applied only to terrorists, the problem is that they're premised on probing all sorts of mainly innocent data transactions, using automated computer algorithms to pick out what seem to be patterns of terrorist activity (but could just as readily be, and would almost always be, completely innocent interactions and relationships).

Unfortunately, many in the establishment seem to agree with the quote attributed to Sun Microsystems CEO Scott McNealy: "You have zero privacy. Get over it." For example, the recent, well-intentioned Markle Foundation report, constructed by luminaries including many with whom I've worked and otherwise respect, seems to have largely caved in on the substance of privacy issues, while paying rhetorical homage to the need to respect privacy and liberty. The report, in several places, praises the initiatives and approach of Poindexter's office, and proposes guidelines for accessing public and private databases that rely on a standard of mere "relevance" to terrorism in the eyes of government officials, whose discretion we're again supposed to trust, and seems to assume without question that even private databases must be open to government searches. This would make McNealy's statement, which was really an overstatement at the time, a reality. At least the Markle report must also be given credit for other language preferring narrower approaches using existing information (e.g. watch lists) to fishing expeditions aimed at gathering as much new data as possible.

"We Are Experiencing Technical Difficulties"

The technical challenges involved with TIA, especially the Genysis centerpiece, are immense. They stem mainly from the sheer ambition of the project and size of the planned virtual database, which compounds the impact of the data quality and interoperability challenges that even ordinary data mining efforts face. There will, of course, be more errors as the quantity of data input from the U.S. and around the world is drastically expanded. Not only could real terrorist activity be harder to detect (the old needle in a haystack problem suddenly becoming one of a needle in an ocean of constantly flowing data); with that much more data flowing through the system, there will likely be many more "false positives" (innocent people erroneously labeled terrorists). To the extent that law enforcement depends in part on legitimacy and cooperation from the population, such a result could not only be of dubious value, but actively counterproductive.

As an example of the sorts of errors that could occur, consider more deeply those credit card fraud detection programs. If you're like me, and travel a lot, you've probably received what seem to be ceaseless calls — "your card was just used in Moscow, a day after it was used in Amarillo" — which, at least in my experience, have never once accurately predicted fraud. This is because the programs rely on automatic analysis of patterns of activity, and are actually pretty rigid. Since some credit card thieves travel around, businesspeople like me are "false positives" and have to put up with all the calls. The system never seems to learn about and adapt to my own patterns, even though I'm a good customer.

In addition to expected errors resulting from the automated systems doing what they're supposed to do, unexpected problems will result from inevitable human error, misuse, and abuse of the information. Recent examples coming to mind include White House Political Director Karl Rove's lost slides providing instructions on how to exploit the war for political gain, or the Navy's loss of a dozen computers containing classified information. These accidents caused embarrassments and are real security breaches. In addition to technical mistakes from the errors in data quality plaguing databases of all kinds — e.g. misspelled names, outdated addresses, unintended variations in formats for data entry — there are common human errors stemming from incorrect interpretations of the data. An example of this is the number of completely innocent people that have missed funerals or pressing business when they've been prevented from flying, because they happen to have names that are similar to those names on the government's controversial "no-fly" list. That list itself illustrates still another example of human error: it has overreached to include not only known terrorists, but also peace activists and some who are guilty of nothing more than dissent against government policies. This, in turn, goes to a core problem with governmental infringements on privacy: they present you out of context, often without any ability to rebut the government position with more accurate, rich details of your own life. We become who the machines, and the government agents interpreting their output, say we are.

In November of last year, an un-manned CIA Predator robotic probe "took out" six suspected al Qaeda members in Yemen, including a U.S. citizen. No evidence was produced, and it is unlikely any ever will be, that confirms any ties between these individuals and al Qaeda, which was one reason some critics classify this as an extra-judicial assassination rather than a targeted killing, legal in wartime. Certainly, the lack of due process raises concerns, as execution is more serious than detention without trial or mere surveillance. This is especially so since the same approach could easily be used within the United States as well as abroad (recall how the military was deployed to assist with the recent sniper attacks in Washington, D.C.). Unmanned flying vehicles like the Predator are shrinking all the time, and reportedly are already being considered for use in the U.S. to help with border patrol and search for suspects. DARPA has already funded programs for nanotechnology "bots" which at insect-size (and approaching molecular size) can individually or in "swarms" monitor and wirelessly communicate about activities, or, in the future, even come together to use collective intelligence to deploy a weapon against foes. (Check out science fiction writer Michael Crichton's new novel, "Prey," which, like most of his work, is based on sound science projected a few years into the future.) Needless to say, the idea of a constant swarm of nanobots joining the increasingly ubiquitous net of audio and visual sensors, GPS location devices, and wireless networks constantly monitoring our actions raises serious questions about technologies concentrating even more centralized power within the Leviathan. The ultimate nightmare would be automated detection technologies hooking up with automated "preemption" technologies to have the war of machines against humans that's at the heart of "The Matrix" films and many other science fiction scenarios.

I expect that some people, upon learning of my concerns, will accuse me of being a Luddite. But the charge has no basis. I've always been an "early adopter" of technologies, am a gadget freak, and have invested in and helped to lead several startup companies based on the sort of pattern recognition, voice recognition, and database and data-mining technologies sought by TIA. These technologies, which include advanced computing methods like artificial intelligence, neural networks, genetic algorithms, and evolutionary computing, can substitute for or provide invaluable software add-ons to the old massive supercomputers or "big iron" previously used to analyze reams of complex data and predict outcomes. My experiences with such technologies has taught me their huge potential in limited contexts, mainly related to such areas as consumer marketing, retailing, and financial services. I'm familiar with behavioral modeling in these contexts, and the power of real-time, dynamic pattern recognition operating on large, complex, non-linear data sets to enhance event prediction and decision optimization (yes, we really talk like this). Frankly, the documents on the TIA website read like a bad dot-com business plan for the anti-terrorist boom times, strikingly similar to the many actual dot-com business plans I've had the pleasure of reading and quite a few in which I regret to say that I invested. (I hope Congress and American taxpayers do not make the same mistake.)

But I'm also acutely aware of the limits to and hype associated with some of these technologies and approaches, and the many flaws in assuming that the technologies can be transplanted wholesale to other contexts and expected to be used effectively against terrorism, for example. Poindexter's response would be that the very nature of the "revolutionary" technologies pursued by TIA and DARPA is to transcend these limits. Maybe . . . but could it be at the cost of ultimately fomenting another American revolution? The challenges of using technology to predict human behavior are toughest of all — even when the humans aren't terrorists trying to outsmart you.

Humans are notoriously complex animals, and unlike most consumers, terrorists know how to and can be expected to adapt, to fit in and hide by behaving like innocent individuals. But while terrorists will continue to do so, the 99.9% of innocent individuals won't evade the system so readily. Pentagon spokeswoman Jan Walker says that "People have to move and plan before committing a terrorist act. Our hypothesis is their planning process has a signature." Poindexter similarly argues that terrorists leave "fingerprints" that TIA aims to detect. The problem is that this fundamental assumption, which underlies the entire TIA project, may simply be untrue. The U.S. experience with terrorism, particularly "global terrorism" of the sort represented by al Qaeda, remains very thin — probably too thin to provide a meaningful model.

A Better Way

After the appalling intelligence failures leading up to Sept. 11, the government clearly needs to do a much better job with the information and resources it has. And certainly the government should not be encumbered by outdated computer systems that don't allow them to perform their essential and legitimate functions. It's far from clear, however, that they need the expanded powers granted them in the Patriot Act and Homeland Security Act, let alone the TIA project. The absence of those powers wasn't the reason for the intelligence failures, and granting the powers endangers the liberties we are supposed to be defending in this "war on terror."

It would have been far preferable to have already had a thorough and independent investigation of "what went wrong" prior to taking all the draconian measures taken, and certainly would have been more conducive to pinpointing and addressing the actual problems. Instead, the Congressional hearings on the matter were delayed and truncated, and the independent investigative commission belatedly established is likely to be no more than another self-justificatory, "feel-good" measure offering merely the illusion that something meaningful is being done.

Intelligence agencies proved themselves grossly incompetent, as they have in other contexts, like evaluating Soviet Union strength or Saddam Hussein's nuclear capabilities.

We certainly had quite a bit of information prior to the September 11 attacks themselves that, properly shared and analyzed, could have tipped off authorities to the planned attack. This included information from the CIA (never transmitted to the FBI) about two of the hijackers who had attended U.S. flight school (Khalid al-Midhar and Nawaf al-Hazmi) and entered the U.S. in early 2001; the FBI's July, 2001 "Phoenix Memo" warning that bin Laden's agents might be training in U.S. flight schools; the Minneapolis FBI's August investigation of Zaccarias Moussaoui for suspicious flight school activities, which was never connected to the previous bits of information; and two Arab language intercepts of al Qaeda transmissions by the National Security Agency from the day before the attacks, indicating "tomorrow" to be "zero day" and the day "the match begins," which were not translated until September 12 due to the paucity of Arabic language translators in U.S. intelligence. Republican Senator Arlen Specter of Pennsylvania, a former prosecutor, has said that in his professional judgment this was not merely a job of "connecting the dots," but a "virtual blueprint" for the attacks that was more than sufficient for the intelligence agencies to have spotted the plot.

Improving the sharing of relevant data between intelligence agencies, and improving their abilities to "mine for gold" in the vast storehouses of data already collected, could clearly have dramatic payoffs in preventing terrorist attacks by focusing on important, quality data. But vastly expanding the scope of data collected to more broadly encompass the activities of innocent Americans or foreigners is a very different thing, worrying even Newt Gingrich, not to mention others from across the political spectrum. Given that the intelligence agencies already receive much more data than they know what to do with (already exceeding each day all the printed pages in the entire Library of Congress), will the solution really be found in dramatically increasing the inflows of that data? This will gather in what everyone acknowledges will be 99.99% irrelevant "noise" instead of the focused "signals" of actual terrorist plots, and is unlikely to help combat terrorism. Rather than exponentially increasing the problem by having TIA looking at the whole universe of all activity by all innocent people, why not focus on terrorists, starting with those on the watch list, i.e. those whose behavior makes them true suspects? Since we know that several of the hijackers were on the terrorist watch list but entered the country nonetheless, isn't it a much more urgent priority to correct the human and technology problems that resulted in such obvious snafus? A more targeted approach would both be more effective in combating terrorism, as well as more protective of our liberties, particularly in this climate of expanding executive discretion and diminishing legislative and judicial protections for those liberties.

Citizen, vs. Government Awareness, Needed

The mind reels at the many ironies in TIA being pushed by the Bush administration, which, as nominally conservative, supposedly believes in limited government, private enterprise, and market-tested solutions. Instead of limited government, TIA is the most intrusive government surveillance project in world history. Instead of supporting the privacy and trust that underlies e-commerce, and the magic worked by private enterprise, TIA would obliterate those values and slap the invisible hand that's feeding it and all (or most) of us. Instead of market-tested solutions, TIA chooses to impose radically intrusive, top-down technological solutions assuming that government planners know best, spending reams of taxpayer dollars to create a deceptive and vicious cycle that legitimates those technologies without evidence and stifles dissent and meaningful discussion and pursuit of more effective approaches. Does it not occur to Bush or the government planners that America's distinctive values, dynamism, and very identity might be gravely harmed by these moves toward a society of such constant surveillance and control?

In a time of such great fear some strong security measures are inevitable, divorced though they might be from genuine security needs or measures that would truly enhance security. But once enacted, they tend to stay. Their impact on culture is often not appreciated. The U.S. is distinguished by our love of individual liberty. To thoughtlessly sweep away that liberty in an illusory quest for more security is to undermine, in fundamental respects, who we are. And this special American identity and culture will hardly be the last, or the only, social cost. One can easily foresee a world in which offhand comments about religion or politics are detected and lead to your being hauled in for interrogation by Ashcroft or his successors. Or a world in which you're forced to smile at home or work because if you don't, the sensors (and censors) will pinpoint you as a malcontent who doesn't adequately appreciate this great country of ours. George's World. (Bush or Orwell? The difference is fading.)

Why haven't more Americans protested these violations of our fundamental freedoms? Part of the reason is simple ignorance: in a climate where even the lawmakers passing these bills admittedly don't read them before they become law, it is tough to expect average citizens to care much. An abstract notion like "privacy" may seem expendable when compared to the alleged tangible anti-terrorism benefits we derive from giving it up. But this underestimates the importance of privacy, which means in practice the autonomy and freedom from government interference that runs throughout all the most important provisions of the Bill of Rights. Part of the reason is also fear, as an extraordinary sense of new vulnerability seems to justify and even demand extraordinary actions. And part of the reason is a pervasive but misplaced confidence that these violations happen only to "others" and that we are not all at risk. Yet without the procedural protections in the Bill of Rights to serve as a check on the quality of decision-making, any of us could easily be wrongly accused without any way to defend ourselves. If history teaches us anything, it teaches that infringements on the liberties of the few quickly become infringements on the liberties of the many. As Justice William O. Douglas said: "We in this country . . . early made the choice — that the dignity and privacy of the individual were worth more to society than an all-powerful police." Whatever technologies or approaches the government adopts should be strenuously tested in light of our values and liberties.

Despite all the 20th-century rhetoric about "totalitarianism," sheer geographic size combined with technological limitations and the complexity of unpredictable human behavior to frustrate the worst ambitions of petty tyrants, though plenty of horrors happened along the way. 21st-century technology, however, for the first time offers a way to transcend the previous physical limits on surveillance by means of the already ubiquitous and ever-expanding electronic networks monitoring our public and private activities, and the new capabilities TIA is pursuing (like voice recognition, biometric identification, automatic translation, and data management) that allow gleaning much more knowledge from that surveillance information. Linking together and mining these data sources on which we depend to communicate, do business, relax, and interact, would make "totalitarian information awareness" a real possibility. It's clear that in the current climate it's not safe to rely on any one branch of government to guard our most basic liberties, but that, as always, strong dissent and citizen action, protest, and vigorous lobbying, is the best defense. Otherwise, like the characters in Orwell's "1984," soon we won't know whether we're watching our televisions, computers, or even the fair blue sky or the artificial lights at night — or whether they're watching and listening to us.

---

*

While his conviction on five felonies (e.g. obstructing justice, destruction of documents, lying to Congress) was later overturned on the technicality that it relied in part on his own immunized testimony, the facts remain that he masterminded the illegal sale of arms to Iran to fund assistance to the Nicaraguan Contras. Both the Iranians and the Contras were considered by our government and human rights groups to be guilty of terrorist acts. So Poindexter — who funded terrorists and provided misinformation about it — is now in charge of using information to fight terrorism. America — what a country!

BACK

© Copyright 2008, Liberty Foundation

Send editorial comments to letters@libertyunbound.com.
All letters to the editor are assumed to be for publication unless otherwise indicated.

Send web site comments to webmaster@libertyunbound.com.