We should soon be able to keep track of most activities on the surface of the earth, day or night, in good weather or bad.
Those words weren't written yesterday by someone at the NSA. They were written by Stansfield Turner, a former head of the CIA, in 1986. They show that electronic tracking – and the hubris that accompanies the possession of high technology – have always been part of the US Intelligence Community.
Edward Snowden’s recent revelations on the PRISM program and the extent of modern surveillance activity by the NSA has sparked a heated debate between security experts and civil liberties advocates who make the moral case that such tracking is incompatible with individual rights in Western democracies.
That debate, however, masks another one, one that is often overlooked but is at least as important: is the utility of such comprehensive tracking being oversold in terms of preserving security? The world is currently enduring a “big data” craze, hoping that an explosion of data will cure a myriad of problems, from preventing terrorist attacks to anticipating the next fad in teen fashion. The craze is dangerous, because if history proves anything, it is that more data by itself doesn't make people smarter, more prepared to adapt or less surprised by change.
Take Pearl Harbor. Why was the US surprised by the Japanese in December, 1941? Was it lack of data? Far from it. In her landmark study of the attack, Roberta Wohlstetter writes “At the time of Pearl Harbor the circumstances of collection in the sense of access to a huge variety of data were (…) close to ideal.” Problems arose, not from too little information, but from the inability to glean ‘information’ from mere ‘data’. The same problem arose before the Cuban Missile Crisis. As evidence mounted that something was going on – in other words, as the data got bigger and bigger– the CIA’s analysts refused to entertain the idea that the Soviets would attempt to put nuclear missiles in Cuba. At the last minute, they were proved wrong, and the Crisis ensued. And the same story played again in 1979 in Iran. There was plenty of data about Iran’s instability, but no one was interested. In fact, Turner wrote the paean to high-technology intelligence collection cited above a mere five years afterthe Iran fiasco had plainly demonstrated that when it comes to strategic surprise, the right questions and analytical focus matter a lot more than the hard data you have. How about 9/11? If FBI headquarters ignored reports from its own field offices about Arab radicals wanting to learn to fly Boeing 767s for no good reason, are we really to believe that 9/11 was a “data problem”? While the 9/11 Commission’s phrase “failure of imagination” is almost tailor-made to stifle curiosity about and diffuse responsibility for the attacks, at least it has the virtue of acknowledging that 9/11 was not a “data problem”. It doesn't seem as if the Boston Bombing earlier in the year was, either.
Business is no different : no industry in history has more hard data than the financial industry. Did Big Data help anticipate the subprime crisis? No. It didn't, because more data doesn't make you smarter or more prescient, as we argued before. The shape of the data mountain always embodies the questions that were relevant to the past. Big data is just a big rear-view mirror, not the clear windshield that IT vendors pretend. It tells us nothing about the future.
Back to the NSA: of course monitoring the Internet and traffic analysis (i.e. analyzing who contacts whom) has its uses. It will catch the occasional amateur terrorist who is dumb enough to write about his project in his emails or to visit jihadist sites to learn how to make explosives, so expect some well-publicized successes in that space. But if the agencies and their customers, i.e. policy makers, believe that mere accumulation of hard data will make America safer, they should think again. As Bruce Schneier, a security expert, argues, terrorism is a needle-in-a-haystack problem, and dumping more hay on the stack isn't going to solve it. In fact, it can make the problem harder. Consider the wonders that modern databases enable: in the summer of 2003, a list of 160 potential terrorist targets was drawn up by the Department of Homeland Security. By late 2003, target numbers had grown to 1,849. By 2005, there were 77,769 possible locations in terrorists’ cross-hairs. Today, there are at least 300,000 such targets in the National Assets Database, including such world-renown events as the Indiana Apple and Pork Festival and the annual Mule Day Parade in Columbia, Tennessee. The Department of Homeland Security is on record saying “The list is a valuable tool”- Imagine the data-mining possibilities! In other words, we gather such data simply because we can, even though such pork-barrel idiocy is a scandalous distraction from actual terrorism prevention analysis and a hindrance to real strategy. In fact, the whole big data approach to security is nothing more than a populist fig-leaf substitute for a counter-terrorist strategy. It’s great, however, if you’re a bureaucrat or politician more worried about the next 9/11 Commission than the next 9/11.
As we document in our book, Constructing Cassandra, an obsessive focus on supposedly hard technical data and big data for its own sake were key enablers of catastrophic intelligence failures at the CIA over four decades. This focus is a legacy of 1950s positivism, the naïve belief that human systems are amenable to Newtonian solutions, and that complex geopolitical situations and social movements can be understood by counting physical devices or parsing Internet log files. They can’t: as the saying goes, God gave Physics the easy problems, and social “science” is a false metaphor.
In other words, most of the media commentary about the NSA programs that Snowden revealed spectacularly misses the point. The real case against these programs is practical: reliance on big data constitutes an illusory technological cure-all peddled by national security quacks. Far from enabling counter terrorist strategy, such technical programs erode civil liberties and cloud the minds of counter-terror strategists. As the nineteenth century diplomat Talleyrand was fond of saying, “It is worse than a crime, it is a blunder”. No volume of data will generate of the right questions or the right analytical focus, so no amount of data will keep America safe, either physically or economically. In fact, the opposite is true: big data is a counter-terrorism placebo. Technology vendors, populist politicians, and ambitious bureaucrats will embrace it. The rest of us should ask many, many hard questions.
