Research

• Studies for Countering Biological Terrorism
• Project Guardian
• Weapons of Mass Destruction (WMD) Detection Study
• Shaping Science and Technology to Serve National Security
• Transitioning DARPA Technologies
• Far Future Warfare
• Division Internships

Intelligence Reform - Seminar 2

Revolution in Intelligence Affairs:

Transforming Intelligence for Emerging Challenges

Need Professionalize and Institutionalize Intelligence

• If intelligence is going to fully emerge as the third American national security profession along with diplomacy and the military service we have got to get to the point where we emulate in some way the creation of structures like the war colleges or like the various research institutes in the Defense Department that allow things like the revolution of military affairs to incubate over time.  

What Should Intelligence Do?

• The American people are not fully comfortable that we do intelligence and all espionage and other things associated with intelligence.  

• If you can fly a satellite some several thousand miles over the surface of the earth and pick up all the signals and all the images that are available to you, we’re fully comfortable with that, as a public.   We are less comfortable with the hands on, yet dirty espionage.  

• Harry Truman had the idea that intelligence had to get above the departmental level and be national in one way or another.   And that was the establishment of the DCI, and the establishment of a national cryptological effort.

• But think about the difference between how we recruit foreign nationals, as opposed to the DEA method, where you try to insert your own people fairly close to the core of the illegal operation.   We may have to come to grips with that.  

• The question of what they want us to do is critical, which leads us to an immediate corollary to this question, what is it they don’t want us to do?   And that has actually a two part question:   What is it they don’t want the intelligence agencies to do, and what is it they don’t want done in their name period.  

• Intel ought to be doing what a newspaper can’t do, not what a newspaper can do sort of, and we’re killing ourselves trying to keep up with the wrong market.  

• What is it that you want intelligence to do that can’t get answered outside, what do you want diplomacy to do.   If you try to ask what do you want intelligence to do outside of a real understanding how you do national security policy, you get into some very, very bad areas.   Did anybody really believe in 47-48 when the combatant commanders were being set up that they would become regional managers with the scope of authority, and then, once you ask that question, do you want the senior American planner/operator in the Pacific to be the commander of the Pacific Command.  

• Do we have the counter intel structure that we want in the FBI doing law enforcement?   In my personal opinion, we won’t have an MI5, but I think the threat of an MI5 is pushing the issue of what the FBI has to be.   And I just can’t imagine the US would ever have a colonial service.  

• What do you want national security to do?   And I think that in the end, what is it that you want intelligence to do, within what national security context?  

• Who do we work for?   Who does American intelligence work for?   And that gets to the whole question of the imbalance if you want to call it that in DoD and non-DoD intelligence that gets to this whole question of tactical vs. strategic and it gets to the changing technical environment we’re in, in which support to military operations, to use a phrase that did not exist in 1952, when NSA opened its doors, could not swamp the national mission at NSA.  

• Think the question we have to look at now is national vs. tactical, is are we seeing the denationalization of certainly cryptology, and some of the imagery.  The last question that’s going to be dealt with in the coming years, beyond what do we do, what don’t we do, who do we work for, is the question of who over sees us.  

• I think in some degree we have always known that production could be annoying or disruptive to analysis.   I’m starting to wake up every day with the fear that production is driving out analysis.  

• Because I think if we get away from production as the thing we do, we will get more to the point of providing service.   And the major service that we provide is the expertise and the analysts themselves.   Its not the type scripts, its not the PDBs, its not even the NIE, it’s the expertise embodied in our workforce… small and expert beats large and inexperienced any day of the week.  

• You can talk to Steve Cambone, you can talk to Mike Hayden, that there will never be enough analysts, if you do the old collect, analyze, disseminate, distribute, you will never have, unless the American people want to solve the national unemployment problem by putting a couple hundred thousand people into the intelligence community, you will never have enough analysts to do that cycle.   And if you talk to General Hayden, one of the things he’s real interested in doing in NSA in the future, is to short circuit that cycle is to send some bit of information at the data level out before it gets processed.   

• Maybe we should allow agencies to buy intelligence, with dollars.   But there doesn’t seem to be any kind of forcing function in the community, because there’s no incentive.   Do what PACOM did a few years ago, and buy it on the open market through open source information.  

• Collection was already overwhelming analysis.   Which brings us to analysis.   In the best of circumstances I would say that we’ve got our hands full, for maybe three reasons.   First, globalization.   No question, we’ve lost our monopoly on information, we’re competing with very high class academics, press, think tanks, we’re responding to a bunch of pundits who know absolutely nothing, but all they have to do is right it in the newspaper, and it will take us a great deal of time to respond.   And we have too much information that we can use easily.  

• We’ve talked a bit about recording vs. analyzing earlier.   I think the community was pretty good at who and what and where, we were worse at how and when, and we were miserable at why.   Unfortunately, that has gotten substantially worse, particularly as we have brought in all these other analysts.   We have a tendency towards worst case analysis.   We’re teaching analysts to go after the worst possible set of assumptions, unfortunately as a result I don’t think that we’re helping the policy planning function.  

• There are certain things that are formally unknowable.   There are questions that are formally knowable, but given the limits on collection and analytic capabilities, they are effectively unknowable.  

• We don’t believe in measuring the means to quantify.   And when we see metrics that measure how well we did in counterterrorism by how many terrorists we capture or kill, we have a real problem with that, because if the only terrorist we capture or kill this year is Osama Bin Laden, I’ll take that as being we had a good year.   If we end up catching fifty or five hundred low level types, there is some value to that, but there is a very real fear that I have that sometimes our business process owners go to far in business processes and have us measure things in quantification that have to be measured in some other way.  

• We may not be definitive, and we may not be in all cases as positive as what we provide to the policy maker, and the policy maker needs to be told that.  

• The DCI must have a weariness equal to that provided to the SecDef.  

• She was very interested in sitting down with us and really try to figure out what our concerns were, which was that we would come up with some kind of internal metric, that we would produce more when you had no reference to external value, which is the real issue here, that any bureaucracy can come up with some internal metric and declare success.    An example I’ve used is this, I’m sure the French Army of 1939 was rated by their Inspector General as being the best Army France has ever had.   It didn’t keep them from getting their clock cleaned in 1940.  

• Why the DoD has so much of the initiative goes back to the role of the combatant commanders.   The SecDef has a go to person for Latin America; the SecDef has a go to person for Asia and the Pacific.   Who is the DCIs equivalent as a go to person?   The answer, in a lot of cases is, he doesn’t have one.   The NIC, the NIOs do some of that, but the reality is that we moved the NIOs more to being, and its an important role, the DCIs entrée into the academic world of expertise, which almost makes it impossible for the same NIOs to be your liaison into the academic world, and to be your go to person for understanding the operation realities within the community.  

• The central agency, whatever that is in a reformed community, is going to have to make a choice, and that is, it can be the hub of the wheel, it can be another spoke in the wheel, it can’t be both.  

• If that organization had worked for the DCI, as the DCI instead of being an office within the DI, the sense that is created is that here’s a proponent of the Directorate of Intelligence at CIA, coming to evaluate NSA’s performance, and is the sense that if you’re within the NSA, you’re going to get a fair review, and I think if you come from the DCI.  

• The only opportunities where we have to train people, and this goes back to the previous question, to think as community officers, is after they’ve been in the business for twenty years.   Boy, that’s the wrong end of the career pipeline.  

• Maybe what Steve Cambone ought to do is taking hiring and personnel assignment authority away from the agencies, and do it centrally from the UnderSecretary of Defense for Intelligence, not so much that you’re going to do it, but you’re going to threaten it.  

• The one piece of the national security toolkit that is in pretty good shape is the military (see Afghanistan and Iraq).   Then you get a catastrophic mess in dealing with the aftermath.  

• I believe that senior analysts at NSA and NGA are every bit as good, big picture and contextually, as those at CIA and DIA, and I think that all source organizations tend to get on a high horse, and I don’t think that’s useful at all.  

• When you look at warfighting demands, for effects based targeting, or for taking down a ship that has a proliferation concern, or eliminating ungoverned areas because of the GWOT, these just aren’t your father’s defense intelligence problems, and I don’t think DoD right now is well postured, so I think we’ve got to take a look at the entire relationship between the DIA and the rest of the services.  

• The CIA is implementing a new compensation reform system; performance objectives; one of the models that they’ve should us was “produced (fill in the bank of) X number of (fill in the blank) whatever it is over a six month period.”   And I’m like “wait a minute, this doesn’t speak to whether or not they were correct, informative, illuminating, advance the knowledge on of an issue in the field.” Now, I crafted my own, that were based upon advanced the knowledge of X in the field by …   But I have no delusions that I’m going to prevail.  

• The impact of what he calls electronic distribution of information and how it fundamentally changes organizations.   And its very tough slugging for a read, but he gets into the distinction between what he calls linear and spatial thinking.   Linear in terms of organizations, without understanding that today’s world is a spatial one, its not necessarily tactical vs. strategic.   And I’m going to use a word that I don’t know the correct meaning of, but I look at it as more situational awareness, and the integration of information in the command element, or the action element, so that information is put to its best use immediately, or at least in as short a time as possible.  

• We spent a decade being fat and happy, and we had a good economy and an administration that wasn’t too concerned about external conditions, now those conditions look a lot more threatening than they did, but they still confuse you.   You don’t have the massive simplification that the Soviet military represented for us during the Cold War.  

• I watched an awful lot of senior policy people play their own analyst.  

• Re:   Horizontal integration, being held up as the answer.   I absolutely disagree, and I’d like to spend a couple minutes on that.   I’m not saying that horizontal integration is bad, I am saying that I do believe its necessary, its just not sufficient.   I do believe there is a bit of a false dichotomy out there, when we tend to think on the one hand of data ownership, on the other, we tend to think of everything to everybody.   Neither one of these models is true, we are somewhere in the middle.   We are always going to be somewhere out of balance, because of the sources and methods issues, because of the US persons sensitivities, there is going to be SAPs, there is going to be operational considerations, there’s going to be that balance to be dealt with, so get over it, recognize, that we’re not going to get everything right.  

• I do think we’re badly suboptimized when we split our analysts, when we have SIGINTers in one building and IA in another building and all source analysts in another building.   I do believe that in virtually everyplace we’ve tried it, when you put them all together, you get a whole that is greater than the sum of its parts.  

• The comment I want to make is that, we’re trapped by our education and our predilections.   We’re incapable of seeing the world as to how it is emerging, except in terms of how we were educated.  

• And these problems occur at three levels.   The first level is at the individual analyst.   There are cognitive failures; you have some stupid analyst, you have people who are too busy.  

• The things that I thought was happening was a lack of curiosity.   He said, I agree with you, but that’s not the analysts’ fault.   It’s the tyranny of the tasking, they’ve got production requirements, they’re too busy to ask questions, so don’t beat them up to get a better image of yourself.  

• Within the IC, there are also issues group work practices and habits.   And one of the things that I keep seeing, is that we’ve built up a serious of processes and analytic habits that work the old problem really well.   They survive exactly because they did work the old problem really.   And what they have become is culture.   And then outside of this are the large organizations, which is where I put the IC structure itself.  

• The second is under recognized misleading wrong answer… given the way most questions get asked, and are then given to the analyst, there’s increasingly little contact between the person asking the question and the analyst directly.   Its not that analyst didn’t understand black and white words, its that the analyst didn’t understand the meta context in which the question was asked.   And therefore has a high probability in being miscalibrated in terms of what isn’t.  

• We built, a system that feeds the information coming into the individual analyst by relevance measures.   And what I can almost guarantee you, and I know people can show, that quote relative materials, is highly redundant, but also enrich what you’re looking for, but also thin out what you’re not looking for, and unfortunately come from the area of where you didn’t quite specify.   So almost by nature, you’ve constructed an information system that helps drive the analyst from the information that might in fact help frame the question.  

• Each of those networks that I’ve described exists in pieces at three different levels.   We need to understand the level.   Are we dealing with a part that’s working on the individual levels?   Are we dealing with the part that has to do with work practice, culture, management, directives, or the overall organization, including budgetary support?   Because if you’re going to fix it, you’d better know where this problem is directed in terms of what is fixed to the part.  

• I have many friends who have left my agency because of it, because the chain of command, this pathology, doesn’t understand what they’re doing, or what their value is to the organization, whereas a peer in another organization would understand and would give them a rating that would be much different.  

• Some of this is organizational, some of this is managerial, for this office, with a real a war going on, with forty to forty five percent of their requirements are generated by telephone taskings, which never shows up on their formal credit assignment sheets.   So they’re getting eaten alive for responding to critical requests that serve as a part of, war is a current intelligence exercise.   Understand that.   It needs to be supported.  

• How do you create these kinds of distributed mechanisms that allow you to work with individuals that aren’t sitting next to you, but you don’t know, because, we went through school together and we went through training together.   And, let me be really explicit about something I’ve gone over earlier.   The craft and guild system relies upon an appropriate balance between apprentices, journeymen, and masters.   What we currently have is a system that is overwhelmed my novices, and a real shortage of journeyman and masters to socialize, train and educate them to help them along.   And that’s one of the ways in which you create a truly vicious circle, particularly given the stress on production.  

• And I think the DI motto today is “I know” and don’t tell me anything else that doesn’t fit in with my DI perspective.  

• When you become an analyst, and you start working in a group, there’s no reference point where you say “well, what do you think N. Korea’s like now?”   All they tell you is “read the last two years of production.”   Another point I’d like to make, in terms of sharing analysts knowledge.   One of the problems of today is that the coordination process occurs and the collaboration and peer review process occurs at the time of the written product.  

• There’s got to be a point where we collaborate about our thinking on the process.   There’s got to be.   But this point doesn’t exist right now in the intel process.  

• You think of your country in those terms, and the information age, and youth politics, and youth culture, religion, demographics, and things that we don’t traditionally track, so they don’t make it to our analysis as objectively as they should.  

• I’d say that the average DI analyst doesn’t understand the difference between complexity theory thinking, quantum based thinking and Newtonian physics based thinking.   They are very intelligent people, but its just not in their field, and its not what they think about.   I think, and Russ mentioned, they tend to think of states as arithmetical.   Countries are non-linear things.   They’ve actually become increasingly non-linear after boundaries have become irrelevant.   So what new analysts are taught is “take ten reports, add them up, subtract two, and voila, you have your conclusion.”   Put that way, it sounds ridiculous, but can anyone argue that’s how intel analysts are taught to do analysis?  

• Why don’t we have real time evaluation?   We have some capability, for while we’re righting on Iraq, or while we’re righting on Iran, to kind of help each other, by assessing it.   It should be part of the process, but it really isn’t.  

• I question whether the personality types and the aptitudes that you’ve been looking for are right.  

• You’ve got the issue of developing a knowledge base.   We’ve been pre-occupied as of late with building a knowledge base on the world of Islam.   Which is a big, and incredibly diverse place, and we’re about as ignorant of it as we were in the early days of the Cold War vis a vis the Soviet Bloc.   Back in those days, the Agency and the IC were active in being funders and participants in this knowledge base building.   In the fifties, we didn’t have much, by the late fifties, we had text books, we had programs in universities, think tanks, we had expertise inside the government and outside the government, and the experts were comparatively indifferent to their organizational location and who did and did not have a clearance.   They were remarkably indifferent to the left right stuff of the McCarthy era, the world of Islam as a knowledge base issue if inundated with ideology, its narrow, few people know anything about it, and they don’t know how to talk to people who don’t know anything about it.   We’ve got a national problem here.   Where is CIA on this?   IC isn’t the only area that needs this help.   DoD, diplomacy, policy, etc.  

• The end of the Cold War wasn’t enough to force a transformation of American intelligence, the budget cuts of the 90s weren’t enough to force a fundamental change of American intelligence, and that 9/11 was not enough to force a fundamental change in American intelligence.  

• If you want the Pentagon to be a regional manager, if you want the Defense Department to be, not just the warfighting, but the peacekeeping and nation building instrument for national security, then fine, but I think when you make that decision, absent a decision, and it just kind of creeps up on you, then you have a problem.   

• We are suffering from more than a decade from national, non-seriousness about national security.  

• In the terrorism context, I believe the community has surged.

• Which vision of globalization do you think is going to prevail?   The upside or the downside, and I think we can differ on where the trends lie, but I think most professional, intelligence community people are arguing that we’re not in the posture to participate in that debate.  

• We’ve never been more in need of a functioning IC, we are woefully unprepared.  

• No one should stop following the Russian strategic missile force, but there are whole new areas of globalization which we are clueless, uneducated about, we have no vocabulary, we are largely unable to follow, and this is to the detriment of our country’s security.  

• And what’s really happened, and XXX put his finger on it, when the community took its cuts in the early 90s, it took its greatest cuts on the journeymen, at the expense of the seniors, they could go elsewhere.   There’s a term that’s become popular in ecology called keystone species.   If you’re familiar with the concept, its not necessarily the species at the top of the food chain, but it’s the species that ties the whole set of niches together.  

• And what we did, throughout the 90s, was literally remove the keystone species from the IC, and I would suggest that this is probably the same case with the DO as in the DI.  

• Globalization has; the world is changing in a way that I could never have imagined, and I just think we have cognitive techniques that don’t work for this world.   You actually mentioned it a couple weeks ago, and I think it’s a great idea.   We need to establish an In-Q-Tel for the social sciences.  

• George Tenet notes that the large structural changes that people are talking about, is that what we really need are more TTICs, more academic centers that look at specific issues.   

• I believe that the interagency model for this problem will be the only way that we will ever get ahead of it, we’ve got more information in one place on the terrorism problem than anyplace in the US government.  

• TTIC is a proxy for the National Security Act, and I think that they went about it in the right way.   I think that if they tried to legislate it, I think that would have been a huge mistake, because we’re allowed to make mistakes, we’re allowed to learn as we go along, and I think that’s important.  

• First, I think that we did a bit of force fitting of analysis into the existing structure, function followed form, as opposed to the other way around, and that was just an absolute fatal mistake.   We started with these notions of a division of labor in the 1990s.   The CIA did political and econ, INR which took it in the shorts with huge cuts, did political, the DIA did military, but then DIA would divide among the services and the commands whole various components of military analysis, and that somehow seemed to work under the assumption that analysis is additive, and that was just a fundamental flaw.   The world’s never worked that way, it never will, but we concocted a recipe for disaster.  

• But, the kinds of challenges the DoD had to face in the 1990s were a mix of political, economic, and socio-religious problems, so we had these wonderful databases that never had any applicability whatsoever.   DIA is now trying to move back away from that, and we are trying to become more sophisticated in our analysis, the irony here is that we’re getting to be more and more like CIA, and I think the country needs to think through that.   Is this a good idea or not?  

• So that I think is the framework for the problems, some quick solutions, these are my suggestions, from a community perspective we need to start with the right set of premises, that the US national security policy is going to be ad hoc very a long time, that we’re never going to see the military technical corps return to the degree of primacy that it once had, hence XXX point that we’re dramatically underfunding the State Department.  

• Someone said that we’ve got to convince our congressional and policy masters of the kind of wisdom of those kinds of things, and I think that gets back to what the DCI said about Centers.   I think in this world, TTIC is exactly the right approach.  

• I think intelligence needs to be close to the consumer.   So I think I want a bit of my China intelligence analysis done by PACOM, because they’re so close to it.  

• Many of in the press are pretty sharing about that for obvious reasons, but I for one believe that we should be identifying who the “good” reporters, press types are, and there should be far more direct dialogue between them and the intel community.  

• My concern with the DNI and similar proposals is that we’re going to go reorganize the boxes or the book without understanding the subtle control mechanisms what the relationships are and the real problems are inside us.  

• We have a vastly diminished competitive advantage in many areas… With the exception of just a few areas, such as counterproliferation and organized crime, I think our competitive advantage is second to many smart think tanks, that have many smart people out there that do better research than many of our analysts.   Policy is certainly less trusting of us.  

• I would argue that means you have got to have a substantial center of mass here in Washington to make a decision about whether you want to participate in that war or not, and that’s going to draw you inextricably into discussions with those regional CINCdoms who are a source of substantial intelligence.   So integration here is key, and I believe that we’ve embarked on a whole series of organizational constructs.  

• What are we doing to redefine the intelligence cycle?   That cycle is about as old as your product, and probably has no value in today’s intelligence community, and I think that’s one of the things that we can start looking at with think tanks, etc., to ask the question, what does the new multi-dimensional, twenty first century paradigm for the intelligence process look like?

• I will guarantee you that our Islamic terrorists specialists at the Agency know nothing anthropologically or socially about young men in the current information feed.   I will guarantee.   We need to stop hiring political scientists, and start hiring cultural anthropologists, but they are absolutely unclearable (laughter, inaudible).  

• Its not the 90 rookie analyst can’t come up with a piece of information, and its not that current reporting doesn’t come up with streams of interesting information, but they’re without context, and its analytic expertise and knowledge that allows those individual pieces to be bit into a larger context that speaks to larger policy issues, as opposed to closing the sensor to shooter issues, which are important, and I’ve worked them since the last Gulf War.   And there are two causes that stem from them.   One is the fact that we don’t do much out here of in depth analytic work inside the community.   Two, the fact that the community is not a discipline.   Tradecraft is used with analysis because it reflects things, and the community is indeed a craft and guild system, and there’s a really good reason for it.  

• The sciences or who have teaching positions, practice their craft, but they also provide building blocks to a whole creed of knowledge, that’s the whole test of a Ph.D:   Did you add something to the field?   In the process of doing that, they grow journeymen analysts.  

• I have a professor who was my dissertation advisor, but in fact, the real learning doesn’t come from him posing the question, it comes from the practice in the lab, and doing the research.  

• It also has two things that the intel community does not have.   It has a formal system of lessons learned that doesn’t occur only after major failures, that transmits itself into formal doctrine that we’re working on.   And an intensive set of exercises and training where training becomes the practice of the craft, inseparable from what you do day to day.  

• We’ve come to the conclusion on our team that what you really want to do is provide appropriate tools for the novice analyst and the appropriate tools for the expert analyst.   They are not the same set of things.  

• There’s nothing that you can do to magically make novice analysts into expert analysts while using a set of tools.  

• We need to speed up, fix it, but I think we’re dealing with three very different timeframes, with fixes unlike Tenet’s testimony, in where it’s a five year problem, and its fixed.   I think there’s a set of short term things, where there’s two years, and we’re done and we’re out.   I think there’s a little bit of five year stuff, and a lot of that has to do with better use of some of the technologies, and its not the volumes of information of stuff flowing in from point A to point B, but flowing it from here to here, helping the analyst do his job better.   And then there’s a fifteen year process, I think we really need to rebuild the journeymen cadre.  

• I floated a proposal five years ago where I floated the idea for basically what became In-Q-Tel, a hardware and analytics skunkworks.  

• The fact that peers tend to be equally prepared in a content area.   The other benefit is really, peer competition.   That is the presumption that there’s a zero sum game.   That is, if Dr. XXXX’s paper gets published, mine won’t.   Therefore, competition is keen; my homework, my preparation is tough, to shoot his stuff down makes him very on his game.  

• The kind of people that we really need and will really need in the future are the kind of people who’ve never thought about working for the CIA.  

• There aren’t that many people out there who have a really a lot of knowledge on these issues, and we need a government program that gets a lot more effort and energy in these academic circles and causes that world to be helpful to the order.  

So just being able to read an academic tome about cultures is fine, but until you are organic in it, do experiments in it, you’re only part of the way there; the academic community can be helpful, but its only one minor part of the way.