In today’s Washington Post, an anonymous intelligence official talked about the intelligence community’s role in the attempted Christmas bombing:
Anyone who believes that a relatively small organization like NCTC [National Counterterrorism Center] is going to connect every electron in each of those 30 databases is either disingenuous or naive, and certainly knows very little about how intelligence analysis actually works.
Bingo! We as a public have to reorient our expectations about the intelligence community’s ability to ensure 100 percent security on a 24/7/365 basis. That’s not a knock on intelligence pros. As a former intel analyst, I’ve seen time and time again just how unrealistic the expectations are.
While individual quotes that dumb down the intelligence community’s capabilities are illustrative, they fail to drive home how difficult intel work really is. I think it’s more useful to examine what it’s actually like to “connect the dots” in the case of a potential terrorist operative. From my own experience, here’s how it works:
I’d receive a lead from the CIA Station in Rabat, Morocco, about a potential bad guy. For the purposes of this explanation, let’s say he’s a Moroccan named Abdul Aziz Mohammed Abu Sayaf, but I don’t know anything else about him, such as his date or place of birth. (I chose that name not because I want to stereotype all “terrorists” as Arab or because he’s an actual bad guy, but because – as I explain in detail below – it will help illustrate a point about transliteration’s role in analysis of suspected terrorists specifically from Muslim countries.)
My goal is to find out everything we know about this individual and determine whether he’s a legitimate threat. This is no small point — in order to raise the alarm, I need definitive intelligence corroboration that the individual in question has a reported history that solidifies him as a potential danger. In other words, we don’t just arrest people because of a single report from a source of unknown quality. For the record, 99 percent of the time, walk-in sources to U.S. Embassies are of poor-to-unknown quality. That includes friends and family members who walk into the embassy and claim their relatives are potential dangers. Why? Family relations are tangled webs, and who really knows if your uncle just might want you arrested in revenge for that unsettled family land dispute.
Therefore, I’ll take his name and plug it into NCTC’s terrorism search, a database that stores more information about terrorism suspects than you could ever imagine. Most of the information is contained in reports from the CIA, NSA, DoD, State Department, and foreign intelligence services that have shared with us. The reports range in length from just a paragraph or two about a specific individual, to tens upon tens of pages long of names, aliases, and birth dates of “suspected” individuals about whom these suspicions are undefined (thank the Italians for this).
“Abdul Aziz Mohammed Abu Sayaf” goes in the old database, and presto-changeo, 27 reports come back. I tear through them for information that matches what I know about my guy. Say I can throw out 22 of the reports because they’re all about an “Abul Aziz Mohammed Abu Sayaf” who lives in Indonesia and was arrested in 2004 and is now in jail.
That leaves five reports. Four are about an Egyptian. Out. And the last one is about some guy of the same name in an unknown country who doesn’t appear to have really done anything wrong. I’m interested in the last one, but need much information on him before taking action.
Here’s where it gets fun. Since there may be more information out there, I start looking for variations of Abul Aziz Mohammed Abu Sayaf’s name, as names like Aziz, Mohammed, and Sayaf can be spelled several different ways when transliterated into English from Arabic. But rather than guess at which combination of the spellings works in our guy’s case, I would enter into the database, “Abdul Aziz* M*h*m*d Abu Say*af*,” which accounts for the different vowels and multiple consonants that may be used in variant spellings.
The result? 2,453 new reports to comb through!
I would logically cut that number down by entering what little other information I know about this guy. Next search: “Abdul Aziz* M*h*m*d Abu Say*af* AND Morocco.” Down to 372. Next search: “Abdul Aziz* M*h*m*d Abu Say*af* AND Morocco adj! 20,” which means all of the above words must appear within 20 words of one another. Down to 87.
I diligently read or skim through all the 87 reports looking for any nugget of information that could corroborate the suspicions about our man. Perhaps I find an additional report or two about an individual who might be the person in question, but I can only say that with 50 percent confidence.
The end result is that I write another report saying only what I can definitively conclude:
Abdul Aziz Mohammed Abu Sayaf is suspected of wanting to enter the United States to conduct a terrorist attack. Sources of unknown quality indicate Abu Sayaf is interested in traveling this month, though it remains unknown whether Abdul Aziz Mohammed Abu Sayaf is a credible threat to the United States.
I file my report, and the receiving officer – given limited resources to follow leads – deems my report interesting, but not urgent.
Two days later, an individual named Abdull-Aziz Muhammad Abou Sayyaff buys a ticket on a flight to Newark and tries to detonate an explosive belt on board. With hindsight, it’s easy to point out the flaws in my analytic process: Should the name spelling be uniform? Why did you limit your search so much? This is national security – you mean to tell me you can’t be bothered to read 327 reports? Shouldn’t we chase down every lead? And etc… sigh.
These are easy and obvious criticisms. And certainly, some improvements can and will continue to be made. However, given the vast amount of American and internationally derived information, the pressing need to run down several searches like this on any work day, and the permanent resource constraints, these are also criticisms by those who don’t understand the tremendous complexity of intelligence work and the diminishing marginal returns of hiring thousands more additional analysts.
In short, finding bad guys is often like looking for grains of sugar on a beach. Unfortunately, we have to accept that we might not find them all.