Did Poindexter Really Believe the Total Information Awareness Program Would Have Protected Our Privacy?

The Watchers

Did Poindexter Really Believe the Total Information Awareness Program Would Have Protected Our Privacy?

The Watchers

Did Poindexter Really Believe the Total Information Awareness Program Would Have Protected Our Privacy?
Arts has moved! You can find new stories here.
New books dissected over email.
Feb. 23 2010 10:02 AM

The Watchers


The Watchers by Shane Harris.

Good morning, Shane,

Why is it that when I picture you and John Poindexter setting sail on the Chesapeake Bay, the Gilligan's Island theme starts playing in my head?


Of course you're right; the book is about not just Poindexter but a series of other scientists, spy bosses, and entrepreneurs who share his fixation with technical solutions to national security problems. My favorite, apart from the ever-slippery Mike Hayden, is Jeff Jonas, the consummate West Coast, high-tech autodidact who developed programs for Vegas casinos to identify card sharks and corrupt dealers, then became a Washington player when the intelligence community realized his innovations might be used to identify terrorists as well.

But as you say, Poindexter is the high priest of these ideas, and I'm glad you brought up the role that privacy protections played in his initial conception of Total Information Awareness. I suppose for Poindexter himself, this must seem like the ultimate bitter irony: It was concerns about privacy that led to his ouster from DARPA, yet what nobody understood was that he, John Poindexter, had been looking to preserve individual privacy by writing privacy protections into the very code of TIA.

This seems to be Poindexter's narrative of events, and it might strike the reader as self-serving, to say the least. But unless I'm misreading The Watchers, it's a characterization you appear to agree with. "If our leaders were truly committed to preventing acts of terrorism, and at the same time really wanted to protect individual privacy," you write, "then they should have kept Poindexter right where he was, as the lead visionary and chief proponent of a radical new way of thinking about how to secure people's lives and their rights."

Here I think we differ. Poindexter's plans to outfit his pervasive spy program with innovative privacy protections are real, of course, and in detailing them, The Watchers complicates the standard Orwellian narrative of Total Information Awareness that many observers (myself included) have perpetuated over the years. But while you complicate that narrative, I wouldn't say that you completely upend it.


Let's look more closely at Poindexter's actual ideas. TIA was designed to aggregate huge volumes of data about people and transactions, then scan for suspicious patterns of activity. If three Muslim men who had overstayed their visas and happened to share an apartment and a bank account all bought one-way tickets on a domestic flight, an alarm would sound, because that pattern has a strong correlation with a terrorist attack. "Now pattern-based searching of everything, of everyone, of all that was known or knowable seemed to [Poindexter] the only logical choice," you write.

To curb the dangers of privacy violations, Poindexter imagined a "privacy appliance," which would encrypt all of the private data about individuals in the system, keeping it "in a kind of electronic safe that might only be opened upon order of a judge. The judge would have to find that the government had a reason for thinking this anonymous person might actually be a terrorist." Poindexter called it "selective revelation," and suggested using the system to track the activity of the very analysts who operated it, creating an immutable audit trail of every search that they conducted, to prevent against abuse.

My first reaction to this suggestion is to marvel at its novelistic perfection: For Poindexter, the tireless innovator and tech-evangelist, the conflict between security and liberty boils down to a mere software glitch, one best resolved not through debate or any form of deliberative government process but through the addition of a few new lines of code. There's a poignant moment when Poindexter consults Marc Rotenberg, of the Electronic Privacy Information Center, and solicits his advice on incorporating privacy protections into TIA. Rotenberg tells the admiral that oversight is important, and Poindexter concludes that the two are on the same page. "[H]e wanted TIA to monitor its own users," you explain. "That was a kind of oversight." Later, when Rotenberg publicly decries the program as a "perfect storm for civil liberties," Poindexter feels betrayed.

My second reaction to Poindexter's privacy proposals is: hogwash. As Rotenberg no doubt would have elaborated, had he thought there was any need, "oversight" entails more than self-policing. And while some privacy advocates might take comfort in the notion of encrypting private information and keeping it in an electronic safe until TIA analysts obtain a judge's permission to look at it, this scenario seems almost laughably out of step with prevailing realities. The Watchers documents in terrific detail the NSA's initiation of a secret warrantless wiretapping program precisely because the traditional requirement of a court order for domestic eavesdropping was perceived as dangerously onerous. You tell a harrowing story about three American soldiers captured by Iraqi insurgents in May 2007 who were killed before U.S. forces could save them, at least in part because lawyers in Washington spent days arguing over whether or not a warrant was required before the NSA could intercept the insurgents' phone calls because some of the calls passed through switches in the United States.

At heart, this issue always boils down to the clash between what technology enables and what our laws allow, and most of the watchers in The Watchers push the technical envelope and chafe at any statutory or constitutional constraint. "You're asking us to take this race car and drive it around the track at twenty-five miles per hour," an exasperated official operating another program objects when the lawyers are slowing him down. My hunch is that most people familiar with TIA, and the program's proponents in particular, might argue that mandating judicial approval before analysts could obtain the actual names of any of individuals targeted by the system would render the whole enterprise impossibly cumbersome.

So I'm left suspecting that Poindexter's vision of a silver bullet that both predicts attacks and protects our privacy amounts either to credulous optimism or to cynical window dressing—or possibly, given the complexity of the admiral, to both.

Or is it me who's the cynic?