Another intelligence failure

Another intelligence failure

by digby

This one is just sad:

Several years ago, a senior officer in the CIA clandestine service attended a closed-door conference for overseas operatives. Speakers included case officers who were working in the manner Hollywood usually portrays spies — out on their own.

Most CIA officers abroad pose as U.S. diplomats. But those given what's called non-official cover are known as NOCs, pronounced "knocks," and they typically pose as business executives. At the forum, the NOCs spoke of their cover jobs, their false identities and measures taken to protect them. Few said much about gathering intelligence.

A colleague passed a caustic note to the senior officer. "Lots of business," it read. "Little espionage."

Twelve years after the CIA began a major push to get its operatives out of embassy cubicles and into foreign universities, businesses and other local perches to collect intelligence on terrorists and rogue nations, the effort has been a disappointment, current and former U.S. officials say. Along with other parts of the CIA, the budget of the so-called Global Deployment Initiative, which covers the NOC program, is now being cut.

"It was a colossal flop," a former senior CIA official said in sentiments echoed by a dozen former colleagues, most of whom spoke on condition of anonymity to discuss a classified program.

The CIA spent at least $3 billion on the program, and the number of specially trained spies grew from dozens to hundreds. The entire clandestine service is believed to total about 5,000 people.

But because of inexperience, bureaucratic hurdles, lack of language skills and other problems, only a few of the deep-cover officers recruited useful intelligence sources, several former officers said.

Apparently, the people of Iran (where most of this was focused) are watching Hollywood movies too and exposed most of these spies who were then brought back to the US.

This is unfortunate. Assuming any powerful nation will spy on other countries, this form of HUMINT is the traditional way it's done, with the use of people with knowledge of the relevant culture and politics using well honed skills to seek specific information on the ground. That's a form of intelligence gathering that relies on the heuristic abilities of human beings which, in my opinion, are a lot more nuanced and sophisticated than the NSA dragnet computer models can possibly be.

But that takes training and long term commitment and we don't really want to bother with that when we can spy on everyone in the world and draw conclusions from inferences based upon who knows who and where they drink their coffee. Let's just say the possibility of error is at least as great with this huge data collection as it is when relying on trained spies on the ground. And I would guess that the possibility of missing something more important is greater. Computer analysis is only as good as what it's programmed to analyze. Human beings on the ground would always have a subtler grasp on the reality of any threat.

But it didn't work out apparently. So we're going with this instead:

Aside from the obvious authoritarian concerns I really have to wonder if reliance on this fancy high tech intelligence gathering is going to end up making us less able to understand the complex nature of what it is we're supposed to be spying on: namely, human beings.