« Intel Gone Bad: Exhib… | Home | Intel Gone Bad: Terro… »

Intel Gone Bad: Exhibit Afghanistan

(On Violence is devoting the month of May--and most of June--to a simple idea with complex ramifications: Intelligence is Evidence. Click here to read the previous posts.)

In American war zones, two different groups use intelligence/evidence to stop insurgents. First, there are the maneuver brigades, battalions and companies attacking insurgent networks through both security and offensive operations.

When it comes to high value targeting, the top leaders and the really bad guys (allegedly), Joint Special Operations Command takes the cake. In Iraq--in addition to the Surge and the Awakening--the increased emphasis on lethal operations by JSOC, led by General McChrystal, contributed to the quelling of violence.

When the American military turned from Iraq and towards Afghanistan, it brought this well-oiled machine with it.  

This machine relies on “intelligence as evidence” to show 1. that insurgent leaders are insurgent leaders and 2. where they are. This machine also wholeheartedly believes in the one percent doctrine. While targeted operations of insurgent leaders are often very successful (U.S. leaders in Afghanistan claim that only 1% of missions involve civilian casualties), when they go bad they go terrible. Here are three examples:

1. "Anatomy of an Afghan War Tragedy". This Los Angeles Times article went, as far as I can tell, unnoticed by the larger milblogging community. The article recreated an attack by a military drone on a convoy of three trucks of civilians. As a case study, it shows what happens when simple suspicion meets the threshold for offensive action.

In short, a Special Forces team conducted a mission in a village in southern Afghanistan. Since SF troops have access to a wealth of support, they had an AC-130 gunship, predator drone and helicopters all in their near vicinity. After spotting two vehicles acting suspiciously--flashing lights at one another in the dark--the team manning the Predator drone determined “hostile intent”. The next few hours were spent following the convoy, then eventually engaging it with Hellfire missiles. Very quickly after the attack, the overhead surveillance realized that the vehicle did not have only “military age males” (a term I wrote about at length before), but many women and children.

"Anatomy of an Afghan War Tragedy" shows what happens when “intelligence” means “less evidence”. In this case, the pilots suspected hostile intent--”intelligence” in the loosest sense--then found the evidence to support their assumption. Reasonable doubt never came up; the pilots simply needed to establish “hostile intent” to their own satisfaction. Could this have ever passed muster in court of law? No way, but this is warfare. Even though the suspect vehicles started driving away from the SF team, the pilots overhead always suspected them of malice. In the end, that is all that mattered.

2. The Case of Zabet Amanullah. I highly recommend the recent FRONTLINE program “Kill/Capture”. It explores the shadowy world of JSOC and its lethal targeting program. It describes the two primary lethal methods of the JSOC: air strikes and night raids. In the case of an air strike on prominent Afghan elder Zabet Amanullah, FRONTLINE and the Afghan Analysts Networks believe U.S. forces killed the wrong person.

Why the JSOC folks hit the wrong person is directly related to intelligence/evidence. Instead of tailing the suspect, for logistical reasons, they relied on “sophisticated overhead sensors”. In this case, the other evidence informing the decision was “precise intelligence” which is as far as the military will admit. Underlying these two narrow intelligence disciplines is the fact that U.S. analysts--who don’t speak Pashtun or Dari and weren’t born in Afghanistan--make these decisions. In short, when JSOC uses less intelligence to prove their case, the odds of an innocent civilian dying rapidly increase.

3. The Detention of an Afghan Elder. So far, I have focused on incidents where U.S. and N.A.T.O. forces kill the wrong people. Sometimes, though, the U.S. military just the detains of the wrong people. A perfect example of this is also in FRONTLINE’s “Kill/Capture” program.

A company of 3rd Brigade 101st Airborne (the Rakkasans, who you might remember from one of our first posts...) received intelligence about a suspected high value target. They go on the mission, but happen to raid the wrong house. Not knowing what to do, and suspecting the village elder of bad intentions (after finding an AK-47 and some magazines), they detain him. On the way, the elder pretty much threatens to never again support the government.

This scene is almost the raison d’etre for this series. It shows the mindset of maneuver commanders on the ground. It is always better to be safe than sorry. I’ll admit this, though: I understand the pressure bearing down on that commander. I have had to make that call before: detain or don’t. In a war zone like Afghanistan the wrong decision can ruin lives. In the long run, it could also ruin our efforts in Afghanistan.

Eric C asked me a pointed question when he read this post, are these anecdotes statistically meaningful? I mean, how do we know that I just didn’t pick and choose anecdotes to prove my point. In fairness, the FRONTLINE folks try to answer this question as well, with evidence both for and against its effectiveness.

In another sense, though, these aren’t isolated incidents. Two incidents in the last few weeks repeated past intelligence mistakes. Like before, civilians were victims of NATO air strikes or night raids, and it caused outrage in Afghanistan. If I wanted to, I could find incidents of collateral damage every month in Afghanistan. Incidents like the third example, where the suspect is wrongfully detained, don’t make the news. How often these incidents happen we’ll probably never know.

I can say, though, that we use “intelligence” to lower the threshold needed for action in a warzone. Partly, this is because we are in a war, not in a criminal problem. Is that the right solution? Next month, I will argue that more certainty is always better, and that our maneuver commanders in Afghanistan could benefit from exercising greater tactical patience.

six comments

I’ve been wanting to comment on the Mike Few piece but haven’t had the time to do it – and myself – justice. In short, what happened to his translator also sent perhaps tens of thousands of Iraqis into the prison system. Think married men with families etc.

This was counterinsurgency meets resistance to occupation. The US. was taking allies from wherever they could find them and if that meant setting former exiles and Kurdish militias against the Arab population, so be it. The Americans could always fall back on the old “ancient sectarian hatreds” canard when they’d triggered a civil war.

Back to topic: I recommend that everyone read the transcript of the conversation in that piece (I read it a few weeks ago). What is important is the need to kill on the part of one or more of the people communicating with each other over the radio. Moreover, it’s worth noting how theirs became the dominant argument. Caution did not win the day because the lives of Afghans are deemed to be of less importance than those of western troops.

@ Steve re: “Caution did not win the day because the lives of Afghans are deemed to be of less importance than those of western troops.”

I, unfortunately, have to agree. There is a lot more to be written on this topic.


There are actually at least four groups competing in the intelligence community (ISAF, State Department, JSOC, and CJSOTF-A). If you include private contractors and other governmental agencies, then you can probably get the number of actors up to ten to twenty. This parochial competition is one of the major problems with the intelligence community. Information is stovepiped vertically, and the actors compete in an internal fight for monopoly of knowledge.

This is nothing new. In WWII, it was the War Department versus State verses the OSS. This political infighting is self-destructive in war. Mainly, because it limits everyone’s understanding of the situation leading to what you call “Intel Gone Bad.”

Information is power; Power for bureaucracies is defined by funding. Until a solution is rendered, then we will continue this cycle of stabbing ourselves in the foot.

Good commanders work through the system through leadership, personal relationships, and sometimes brute force to steer the groups towards mission accomplishment over bureaucratic drift.

The classic case study is Graham Allison’s Essence of Decision. Allison studies the Cuban Missile Crisis to try and determine why Kennedy could not get the proper information in a crisis.

All the while creating accidental guerrillas .


I read this blog from time to time, and thought this was a really good article. I haven’t seen the “Kill/Capture” special but would like to comment on a few things. I’m not sure if I would consider night raids a “lethal method”, more often than not, they end up with not a single shot fired. The fact that operators do not speak Pashto is almost irrelevant, that is why we have maneuver units that conduct COIN on a daily basis. JSOC is really good at what they do (manhunting), and the effects they have on a network far outweighs the occasional mishap because for every one instance of a CF mistake, we can point to hundreds of insurgents caused civilian casualties. I think kill/capture works, you’re crippling the network to a point that they’re willing to negotiate which is what we eventually hope to accomplish right?