Chaotic Dynamics: A Novel Approach to Intelligence Analysis in Asymmetric Warfare

[JeffC]

As stated, mirror imaging, along with falling into logical fallacies and permitting various biases to determine judgments, are a common failing among new and poorly trained analysts. However, "common" does not equate to "pervasive".

If what you're saying is correct, then that means that the CIA is filled with incompetent senior people. What other explanation can there be for not implementing proper training for new analysts year after year?

For the record, I don't believe that to be the case. I just don't have another explanation for how this problem can properly be attributed to habitual poor training.

There are also significant differences in quality and type of analysis produced by the different agencies and services. There is perhaps an agency or two out there that hires far too many people who don't really deserve the title of "intelligence analyst"; but there are others that are lucky enough to have several consummate professionals on the payroll. Throughout the community we have many outstanding analysts who are very much in touch with reality as it exists on the ground and in the heads of the threat.

I think so as well.

Reads the literature? Well, I've spent a couple of years as a collector and worked a bit of analysis here and there, and, as limited as my academic qualifications may be, I stand solidly by my opinions.

That's understandable.

The disparagement of seemingly simple solutions is in itself an analytic failing. Despite the apparent simplicity, they are still dependent upon the human vagaries of the people implementing them. The existing problems with selection, training and professional development of career analysts may have a simple solution - but the primary obstacle is not so simple, as it tends to lie with bureaucrats outside the field of intelligence analysis. If you were able to conduct such a study in-depth, you would find that the degree of severity of these quality problems ebb and flow as much with changes in senior management as it does with hiring and training of analysts.

I don't deny that it's a complex problem, nor that there can be a multitude of possible solutions. Mine is simply one of many. As long as solutions are being offered, implemented, and evaluated, I think that's all that anyone can expect, particularly in very large organizations.

Along the lines of my last statement, I recommend giving a read of Knowing One’s Enemies – Intelligence Assessment Before the Two World Wars. The book is a collection of essays that amply illustrate the point that, even when a nation is in possession of sufficient intelligence of a quality to make effective policy decisions, it can all come to disaster due to the inherent biases, proclivities and abilities of key policy makers. The harmful effects of internal disputes within intelligence agencies, and turf battles between competing agencies, are also laid out in careful detail.

Thanks for the recommendation. I'll add it to my reading list.

This is where you complexity in reaching an effective solution. Hell, its been over six years since the wake-up call and we're still not ready for work yet. Hiring, training and developing good analyst is the easy part.

Don't get me started.

[Ken White]

I think he is correct.

However, I do not think your analysis of his statement is necessarily correct:

"If what you're saying is correct, then that means that the CIA is filled with incompetent senior people. What other explanation can there be for not implementing proper training for new analysts year after year?"

It is entirely possible that said senior people are competent but either bureaucratically constrained and more concerned with protecting their institution or simply are strongly inclined to not stick their necks out to demand competence and release those who are not capable.

Those alternatives are similar but not synonymous though the result is the same. There is, of course the possibility that pure incompetence is part of the problem and there are other alternatives but experience with our government leads me to believe my provided alternatives are more probable; as they say, "indications lead me to believe" the bureaucracy and it's natural ally, risk avoidance, are your culprits.

For the record, I don't believe that to be the case. I just don't have another explanation for how this problem can properly be attributed to habitual poor training.

Having briefly been an analyst -- and not a good one, I'm too impatient and arbitrary; realized that and went on to other things -- I have to agree with others who point out that we can produce a large number of fair analysts but that good one are naturals and hard to find. I'm not at all sure they can be created without that talent. Lot of good basketball players out there, very few Michael Jordans.

I have been immersed in another culture fairly thoroughly for over a year -- I think maybe the Jesuits have it right; get 'em before they're seven -- after that, the culture in which they got to that age is pretty thoroughly embedded and is unlikely to be shed.

There are always exceptions but I'm unsure how you'd determine who would be receptive and who not.

You also have the factor that thorough cultural attunement with one grouping does not provide better analysis capability for a multinational and multi ethnic grouping...

Just some thoughts for your consideration.

Regards,
Ken

[Shivan]

For example, a Red Team of intelligence analysts are deployed to a facility where they are immersed in an environment as realistically constructed as possible to simulate the daily experience of a member of Al-Qae’da in Iraq, or the Taliban in Afghanistan. This may include religious education in the Koran, daily prayers to Allah, all written and spoken communication done only in Arabic, and other like-minded activities. Any scenario that displaces the previous mindset of the analyst and replaces it with one more closely aligned with the terrorist would accomplish the goal.

Usually, one is a native speaker of only one language. Native speakers of Iraqi Arabic are English As A Second Language (ESL) speakers. They will have the same kinds of cultural disconnects when trying to get the words right for their English native speaking employers. And, that is without consideration for such things as idiolects, regionalisms, colloquialisms, and slang.

Interesting. The problems I see with simulation as described above is that we cannot replicate the social-cultural and religious norms and rituals learned natively, beginning in the crib; nor the language issues that "wm" raises above. One of my Arabic instructors replied to a question I posed, "the only way you can learn [fill in the blank] is if you have an Arab mother." Holds true for different situations, I have discovered. Further, many Arab relationships are via kinship ties, which prevent us from truly entering their domains socially, or being seeing them with their guard down, i.e., with their true souls bared.

Simulation/immersion is great, and will go a long way toward reducing mirror imaging and cultural ignorance, but it can only do so much.

Thanks all for your comments. Very helpful.

[nichols]

Simulation/immersion is great, and will go a long way toward reducing mirror imaging and cultural ignorance, but it can only do so much.

We are striving to identify as much of the cultural ignorance as possible during future mission rehearsals using simulations. Multiple stovepipes are being worked on now which will eventually within 12-18 months lead to a mission rehearsal capability that should cover as much of the warfighters environmental spectrum as possible.

The foundation of the simulation is the VBS-2 / VTK:

http://www.vbs2.com/site/index.html

http://community.bistudio.com/wiki/Virtual_Tool_Kit

The requirements for the VTK were driven by a cognitive task analysis done for anti terrorism force protection. The end state for the VTK is a tool kit that is as 'nicholized' as possible in order for the non-gamer trainer to create realistic scenarios with geo-specific virtual terrain. A key piece is the open API that will allow the Marines to 'plug-in' third party artificial intelligence.

The next stove pipe is the Tactical Language series of simulations that currently uses the Unreal game engine:

www.tacticallanguage.com

This contract is known as ALTS; Automated Language Training System. A portion of it is plugable language and geo-specific culture artificial intelligence that will give the trainer the ability to verbally engage the avatars in the target language. Currently it is being built to plug into VBS2/VTK and DARPA's Real World simulation. The other portion of ALTS is a web based application that we are doing in conjunction with the Air Force (gotta aim high). This will be available through Marine Net as part of distance education.

The next stove pipe is part of Doc Silverman's PMServ artificial intelligence:

http://www.seas.upenn.edu/~barryg/HBMR.html

ONR is working on this one, eventually it will get transitioned over to Trasys.

We are attempting to develop a capability to do the following scenario:

The MEU is sitting off the coast of Africa, Burkina Faso has just launched the balloon and a NEO Op is ordered. On the big boat the intell bubbas pull down NGA data sets for Ouagadougou and import then into the virtual world. Mean while back at the embassy the Gunney or Embassy Stan type bubba sends the detailed buildings that they have created of thier hooch via secure line to the big boat. The intell bubbas put this into the virtual world. Now they pull up the AI library and insert Sahel French language and culture plugin. The final effort is edit the artificial intelligence to match current intel. Now the intell bubbas can go watch a movie.

Down in the bowels of the ships, the running around the flight deck 300 times to get a miles lenth of running has ceased. No more chow lines, the Marines are plugging the DVTE computers into the SWAN to begin mission rehearsal for the upcoming NEO. During the rehearsal da Gunney notices that everytime Rudy with the rusty rifle in the third rank that never gets the word...talks to a local avatar, the avatar gets pissed off. A decision is made to assign Rudy to LZ security and he is not to talk to any local people. After the NEO has finished, Rudy is told in loving terms to log on Marine Net and take the targeted language course because they still have another 5 months to go before they are CONUS.

We aren't to far from being able to create the above scenario.

[Stan]

Hi Shivan !

Interesting. The problems I see with simulation as described above is that we cannot replicate the social-cultural and religious norms and rituals learned natively, beginning in the crib; nor the language issues that "wm" raises above. One of my Arabic instructors replied to a question I posed, "the only way you can learn [fill in the blank] is if you have an Arab mother." Holds true for different situations, I have discovered. Further, many Arab relationships are via kinship ties, which prevent us from truly entering their domains socially, or being seeing them with their guard down, i.e., with their true souls bared.

Simulation/immersion is great, and will go a long way toward reducing mirror imaging and cultural ignorance, but it can only do so much.

Thanks all for your comments. Very helpful.

I don’t think we’ll ever get that many soldiers to ‘natively’ learn the Arabic language, cultural and religious norms. Even though I’ve lived and worked here for 12 years, speak colloquial Estonian and even recite poetry from the 13th century, I still can’t fool everybody. You’re 110% correct, you have to start from birth.

But that’s not the point of the simulation training packages today. I think Jeff's first post explained it best.

The first step is realizing that what we think we know about any situation is several steps removed from that situation’s authentic components, based on how the brain processes information coming from outside of itself. The solution is to consciously break down and rebuild the brain’s internal model of the environment in question (i.e., the cultural and societal influences of religious terrorists in the Middle East or Central Asia).

This will in effect allow common soldiers and analysts to not only understand the Arabs (or Africans) better, but more importantly those nuances such as social taboos. Although I’m very effective on my own, it took years to get ‘there’ and I had nothing more than 13 weeks of language training to prepare for my tour. That barely prepared me for anything Estonian.

Paul and I (we were both previously station in Africa) have observed how a typical military situation can rapidly escalate from ‘calm to a firefight’ over little more than a misunderstanding. One could simply say that 50% of the blame was the African that didn’t understand the circumstances or situation, and fair enough.

But what we’d like out of our soldiers, is to remove the other 50% and level the playing field. The other good news about simulation training is getting shot, does not result in death

[wm]

This will in effect allow common soldiers and analysts to not only understand the Arabs (or Africans) better, but more importantly those nuances such as social taboos. Although I’m very effective on my own, it took years to get ‘there’ and I had nothing more than 13 weeks of language training to prepare for my tour. That barely prepared me for anything Estonian.

Paul and I (we were both previously station in Africa) have observed how a typical military situation can rapidly escalate from ‘calm to a firefight’ over little more than a misunderstanding. One could simply say that 50% of the blame was the African that didn’t understand the circumstances or situation, and fair enough.

But what we’d like out of our soldiers, is to remove the other 50% and level the playing field. The other good news about simulation training is getting shot, does not result in death

I have these visions of a scene in the movie "Men In Black," the one where during marksmanship testing with the military's "best of the best," Will Smith/James Edwards blows away the sim of the little girl carrying the physics book rather than the aliens being blasted by everyone else. I would be very impressed if our situatonal awareness training would allow a successful graduate to be able to give the kind of explanation that Smith gave for his action:

[in a shooting range, confronted with numerous menacing-looking targets, Edwards shoots a cardboard little girl]
Zed: May I ask why you felt little Tiffany deserved to die?
James Edwards: Well, she was the only one that actually seemed dangerous at the time, sir.
Zed: How'd you come to that conclusion?
James Edwards: Well, first I was gonna pop this guy hanging from the street light, and I realized, y'know, he's just working out. I mean, how would I feel if somebody come runnin' in the gym and bust me in my ass while I'm on the treadmill? Then I saw this snarling beast guy, and I noticed he had a tissue in his hand, and I'm realizing, y'know, he's not snarling, he's sneezing. Y'know, ain't no real threat there. Then I saw little Tiffany. I'm thinking, y'know, eight-year-old white girl, middle of the ghetto, bunch of monsters, this time of night with quantum physics books? She about to start some ####, Zed. She's about eight years old, those books are WAY too advanced for her. If you ask me, I'd say she's up to something. ...

[JeffC]

Interesting. The problems I see with simulation as described above is that we cannot replicate the social-cultural and religious norms and rituals learned natively, beginning in the crib; nor the language issues that "wm" raises above. ...

Simulation/immersion is great, and will go a long way toward reducing mirror imaging and cultural ignorance, but it can only do so much.

Agreed. One of the most surprising revelations to me as I was working on that essay was the artificial roadblocks that CIA has in place in terms of their hiring practices. Lack of qualified translators is a serious issue for them, yet they reject applicants who are, in fact, from the Middle East because they don't have a four-year degree, or because they're gay (which is even more ludicrous).

[Brian Hanley]

I was a bit of a leader in VR once, have a videotape "Fundamentals of Virtual Reality" sold to over 4000 universities around the world through UVC (now defunct.) Worked on image processing real time acquisitions since. I am aware of the state of AI etcetera. We are a ways from that scenario still.

What we aren't so far from is a situation where actors can play virtual parts who are in a distant location. That's doable now.

That said, while military regulations may restrict fraternization, enforcing those for intel gathering is just plain silliness. Those are the networks that get you places.

For instance, staying out of the drug transport/payment networks means you are kept away from Osama's support system. Get into it in Afghanistan, you'll have a shot at finding him. Staying out of gun running networks means you can't keep tabs on who is selling what to whom without a scorecard. Etcetera.

[Ken White]

...

That said, while military regulations may restrict fraternization, enforcing those for intel gathering is just plain silliness. Those are the networks that get you places.

For instance, staying out of the drug transport/payment networks means you are kept away from Osama's support system. Get into it in Afghanistan, you'll have a shot at finding him. Staying out of gun running networks means you can't keep tabs on who is selling what to whom without a scorecard. Etcetera.

Now, if we can just convince those great brains in Congress who insist on no contacts with the less savory much less the totally illegal...

[Jedburgh]

....while military regulations may restrict fraternization, enforcing those for intel gathering is just plain silliness. Those are the networks that get you places.

For instance, staying out of the drug transport/payment networks means you are kept away from Osama's support system. Get into it in Afghanistan, you'll have a shot at finding him. Staying out of gun running networks means you can't keep tabs on who is selling what to whom without a scorecard.

There is a lot of opinion based on assumption in this thread; much of that assumption itself based on poorly informed and/or dated sources.

While the nuts and bolts of source ops and building networks is not something that is up for discussion in this forum, I will just throw a thought out there for y'all:

To infil Afghan, Albanian, Indonesian, Nigerian or Russian organized crime networks - let alone getting into the core of al-Qa'ida, Chechen or JI terror structures - ain't as simple as just getting someone to approve mixing in with the bad guys - or of the operator being up to speed on language and cultural nuances. These groups operate on tight-knit trust networks, and are generally composed of individuals linked by kinship and friendship ties that are very difficult to compose and live cover to enough of a degree to get inside. It is also not exactly easy to spot and recruit someone inside one of those targets from the outside. So its a helluva difficult and dangerous job that should never be underestimated or taken for granted - but don't assume that nothing is going on in that regard.

[Shivan]

Thank all for the comments.

. . .We aren't to far from being able to create the above scenario.

This stuff is amazing. Thank you for posting the links, esp. to Silverman's work at U/Penn. I had no idea how close simulation was to the real deal. I'm still reading through material to educate myself.

To infil Afghan, Albanian, Indonesian, Nigerian or Russian organized crime networks - let alone getting into the core of al-Qa'ida, Chechen or JI terror structures - ain't as simple as just getting someone to approve mixing in with the bad guys - or of the operator being up to speed on language and cultural nuances. These groups operate on tight-knit trust networks, and are generally composed of individuals linked by kinship and friendship ties that are very difficult to compose and live cover to enough of a degree to get inside. It is also not exactly easy to spot and recruit someone inside one of those targets from the outside. So its a helluva difficult and dangerous job that should never be underestimated or taken for granted - but don't assume that nothing is going on in that regard.

Agree wholeheartedly, which is what I was thinking of vis-a-vis simulation and immersion for intel collectors. Useful, but hard to break into the kinship networks. The old Italian mob started getting infiltrated as the old generations moved on, and the American born crowd allowed people in who did not have ties to the old country; this is happening a little with the Russian mob in the U.S. Long story short - hard to do with AQ etc, but not impossible, as Jedburgh notes.

For intel analysts, simulation & immersion will improve their skills exponentially.

Salaam y'all

[nichols]

Brian,

In your attempt to be the Renaissance man of the board, you are exposing your ignorance on certain subjects.

I even provided links to the multiple thrusts to this project.

I was a bit of a leader in VR once, have a videotape "Fundamentals of Virtual Reality" sold to over 4000 universities around the world through UVC (now defunct.) .

The operative words are 'once' and 'defunct.' This helps explain why you don't work at DARPA, ONR, or answer our BAAs and FedBiz Ops announcements.

I wrote a Workbook for a simulation that is still used for the SSgt Non-resident Course PME and for the FA-57 course out at Leavenworth. That's past history, granted it got me this job 5 years ago but the knowledge, while a cutting edge application for a simple simulation was big news back then....it's dated.

Worked on image processing real time acquisitions since. I am aware of the state of AI etcetera. We are a ways from that scenario still.

Try reading the links, we are within 12-18 months to having this capability. The pieces are already developed, we have tested them using HLA, now we are doing direct plugins using open APIs.

All of this has been made possible through multiple cognitive task analysis, training effectiveness evaluations and pushing the limits on industry.....it ain't rocket science.

I've been in the acquisition side since Novemeber 2002 ; I'm still getting paychecks, I must be doing something right.

What we aren't so far from is a situation where actors can play virtual parts who are in a distant location. That's doable now.

Brian, this has been going onin the civilian side for at least 10 years. On the military side for over 6 years. Fort Lewis uses units in country to send back and train TTPs with CONUS units. In July we had Marines at 29 Palms in a convoy trainer calling in AC-130s in the simulation at Hurlbert using the J-10.

That said, while military regulations may restrict fraternization

I missed that memo, I was 'fraternizing' while in country and married her 3 years later in CA......guess I'm a bad boy, shame on me.

[Brian Hanley]

Brian,
In your attempt to be the Renaissance man of the board, you are exposing your ignorance on certain subjects.
I even provided links to the multiple thrusts to this project.

Looked at them. Careful about assumptions regarding responses. I developed the virtual work immersion system design which is still, today, ahead of its time. (BTW - UVC was a distribution outfit, not an engineering firm - I mentioned that for completeness if anyone tried to look it up.) I solved a basic problem of synchronization unique to battlefield sims (bullets move fast) that is doubtless still in use. (Or it should be. Things sometimes get lost.) I'm the kind of guy that is sometimes sitting there while some board chairman or stuffed suit blathers bull#### to people who don't know any better. Always remember that a demo is just that, a demo.

... we are within 12-18 months to having this capability. The pieces are already developed, we have tested them using HLA, now we are doing direct plugins using open APIs.

What is "this capability"? There's a long ways between "pieces developed", APIs and a VTK and putting high level stuff together.

I've been in the acquisition side since Novemeber 2002 ; I'm still getting paychecks, I must be doing something right.

I'll note that I have been on the development side for considerably longer than that, and that the military has been supporting this stuff for 19 years. (With similar predictions from folks in your position along much of the way.)

Let's go through, point by point, your statement that generated my response.

The MEU is sitting off the coast of Africa, Burkina Faso has just launched the balloon and a NEO Op is ordered. On the big boat the intell bubbas pull down NGA data sets for Ouagadougou and import then into the virtual world.

Completely doable. Those can be put together and canned ahead of time.

Mean while back at the embassy the Gunney or Embassy Stan type bubba sends the detailed buildings that they have created of thier hooch via secure line to the big boat. The intell bubbas put this into the virtual world.

Doable. But I think it's unlikely those building plans would come from the embassy unless they had them on file. It would make far more sense to have them on file stateside. For more likely you'd have satellite photographs and stateside (or shipside) teams putting together any custom VR world than to offload it to somebody at an embassy. Even definition files can be tricky.

Now they pull up the AI library and insert Sahel French language and culture plugin. The final effort is edit the artificial intelligence to match current intel.

You could make some very dumb bots and place them in the world. You might even be able to do some degree of speech recognition. You could have a library of canned "characters" but they would not be capable of intelligent interaction. Look, I am quite aware of the state of "AI", it's various branches, production systems, (Minsky-esque) neural nets, KNN, Fuzzy systems, etcetera. Whether "AI" even exists is highly debatable. What we have now can, if stretched, imitate a cockroach's intelligence. We aren't up to mouse yet. Everything else is just a production system or an illusion of grammar, a la ELISA.

Which is not to say such things can't be useful as far as they go. But be careful about overselling them and particularly careful about throwing "AI" around. Cheers.

[Rex Brynen]

...or because they're gay (which is even more ludicrous).

The CIA ended its prohibition on hiring gay employees in 1991. Indeed, there is even an Agency Network of Gay and Lesbian Employees.

[Stan]

The CIA ended its prohibition on hiring gay employees in 1991. Indeed, there is even an Agency Network of Gay and Lesbian Employees.

Thanks for that, Rex.

I think Jeff needs real CIA contact and perhaps even an interview. Just because an individual claims to be an area expert or linguist, does not automatically equate to employment. There's a lot of gays around today...just don't tell, and I won't ask

The general thrust of this thread was a mistake an (ahem) agent made by watching a video, but yet missed something (I feel he/she fell asleep during that exciting video). We then concluded his/her brain was shut down or poorly programmed. Huh ? If one has the attention span of a gerbil, there's little modern science or the CIA can do...it's too late.

[JeffC]

The CIA ended its prohibition on hiring gay employees in 1991. Indeed, there is even an Agency Network of Gay and Lesbian Employees.

Indeed. My mistake. It's the Defense Language Institute that gave the boot to Arabic linguists for their sexual orientation (and still will).

CIA still has internal roadblocks (like a 4-year degree among other issues) that are stopping them from filling a critical need for qualified translators. As of June, 2005, they are supposed to be reviewing those procedures.

[Jedburgh]

Indeed. My mistake. It's the Defense Language Institute that gave the boot to Arabic linguists for their sexual orientation (and still will).

Just to be clear, its not DLI that kicks out service members for their sexual orientation, it is the individual service commands at DLI. Speaking to the Army side of that, there are only two ways in which a soldier can be booted out for their orientation: be caught in a homosexual act or put in writing that he (or she) is a homosexual and run it up through a bureaucratic process confirming their sexual orientation several times prior to the actual chapter taking place. A soldier has to work at it to get kicked out for being gay.

During my time at DLI, the soldiers I observed being given the boot were consciously trying to get out of the Army. They fell into two categories. The first were the slackers and cowards, who decided that Army life was not for them (or the thought of being deployed into a combat zone after graduation scared the hell out of'em), and that declaring homosexuality was an easy out. Still a lot of red tape, but faster than the old "food-for-freedom" route (which was eventually closed).

The others were the parasites - they'd wait until they were about to graduate from the course, and then declare their homosexuality and get themselves kicked out after having completed nearly two years of world-class language training at taxpayer expense.

However, taken as a whole, the numbers kicked out for this reason were insignificant compared to either disciplinary chapters or academic attrition.

[JeffC]

For example, a Red Team of intelligence analysts are deployed to a facility where they are immersed in an environment as realistically constructed as possible to simulate the daily experience of a member of Al-Qae’da in Iraq, or the Taliban in Afghanistan. This may include religious education in the Koran, daily prayers to Allah, all written and spoken communication done only in Arabic, and other like-minded activities. Any scenario that displaces the previous mindset of the analyst and replaces it with one more closely aligned with the terrorist would accomplish the goal.

Once the internal structures of the brain have been dissolved and rebuilt, these analysts will be able to discern and predict enemy movements and plans from a much deeper and richer foundation of knowledge than ever before. While the axiom that “we don’t know what we don’t know” will still apply, these analysts will be much better prepared to spot the guerilla in the center of the basketball court.

It's nice to see that Bruce Hoffman agrees with me regarding the need to get into the enemy's mindset. Here's a quote from his article "Countering Terrorist Use of the Web as a Weapon" in the 12/2007 issue of CTC Sentinel (Thanks, Ted)

"To do so, we first need to better understand the mindset and minutia of the al-Qa`ida movement, the animosity and arguments that underpin it and indeed the regions of the world from which its struggle emanated and upon which its hungry gaze still rests. Without knowing our enemy we cannot successfully penetrate their cells; we cannot knowledgeably sow discord and dissension in their ranks and thus weaken them from within; we cannot effectively counter their propaganda and messages of hate and clarion calls to violence; and, we cannot fulfill the most basic requirements of an effective counter-terrorist strategy: preempting and preventing terrorist operations and deterring their attacks.

"Until we recognize the importance of this vital prerequisite, America will remain perennially on the defensive: inherently reactive rather than proactive, deprived of the capacity to recognize, much less anticipate, important changes in our enemy’s modus operandi, recruitment and targeting."