2004 Robert Steele: Reinventing Intelligence, From Truth, Power




President, OPEN SOURCE SOLUTIONS, Inc. Editor’s Note: [Mr. Steele is a veteran of the clandestine service with three back-to-back tours in Latin America]. He received two awards while an operations officer. He has also served in offices responsible for programming satellite systems and advanced information technology applications, and concluded his government career as the senior civilian responsible for standing up the $20 million Marine Corps Intelligence Center, and subsequently as a senior civilian in the C4I Department at HQMC responsible for National Foreign Intelligence Program and General Defense Intelligence Program resources under Marine Corps cognizance. Sometimes controversial, always visionary, he is one of those influencing the future of our intelligence community.

The time has come to outline a vision of the future of American intelligence. It was distressing to me, in reading Alvin and Heidi Toffler’s latest book, War and Anti-War: Survival at the Dawn of the 21st Century, to discover in their chapter on “The Future of the Spy” that I am not only considered the “rival store” because of my emphasis on the importance of open sources of intelligence, but that I am considered (by the Tofflers, who have become good friends) to be “the single most forceful enemy of secrecy in Washington”. The Tofflers are brilliant, and I am indebted to them for focusing attention on these issues, but I would like to tell my peers and colleagues in AFIO where I really stand, and in this way solicit comment. All of us, retired or serving, must come together to reinvent ourselves.

I’ll sum up my vision in a nutshell: let’s get back to basics by doubling the clandestine service, refocusing our technical investments, dramatically increasing our consumer base and the voice of our consumers in validating collection and production, and by leveraging–for the first time–the vast “virtual” intelligence community that exists outside the beltway. At the same time, we can significantly improve the totality of our government’s ability to capture unclassified information critical to national competitiveness [by realigning positions within our Embassies to increase the effectiveness of State and Commerce as collectors of unclassified information.] If the intelligence community is to remain a significant element of our national power, then it must recognize that its present form is destructive and counterproductive. The intelligence community cannot and will not survive at its present funding levels unless radical changes are instituted immediately.

Let’s begin with open source intelligence (OSCINT), since that it the arena where I first confronted our inadequacies. When Colonel Walter Breede and I first stood up the Marine Corps Intelligence Center in 1988, with $20 million in funding, courtesy of the General Defense Intelligence Program, it never occurred to us to question the value of the existing Top Secret/Codeword architecture and databases, and so we set in motion a spending plan that ultimately led to $10 million being obligated for a classified communications and computing system with direct access to CIA, NSA, and DIA data. Since Walt and I had always been collectors, it just blew us away when we went to produce intelligence and our analysts discovered that the Third World “data fill” has simply not been accomplished over the years. We are paying the price today for our obsession with the Soviet Union.

The good news is, as I discovered when the shock wore off, that there is an enormous “virtual” intelligence community out there, outside our vaults, which has much of the encyclopedic information that we require to satisfy the needs of our traditional consumers, as well as the needs of new consumers such as the Environmental Protection Agency. It came as a surprise to me, when I devised the first open source exploitation strategy for the Marine Corps, to discover that I could meet most of our needs by spending $20,000 a year on commercial online services such as LEXIS/NEXIS (Mead Data Central), DIALOG, BRS, and the U.S. Naval Institute’s PERISCOPE, plus a few CD-ROMS such as those available from Jane’s Information Group. The first strategic intelligence product of the Marine Corps Intelligence Center, a study of 69 geographic entities, was completely unclassified and based solely on unclassified sources.

The bad news is that neither our architectures nor our procurement practices lend themselves to exploiting open sources. Our architectures, which meet security standards for Top Secret/Codeword, essentially comprise a cement bunker without windows, a bunker into which the fresh air of open sources does not readily enter. Our procurement practices are overwhelmingly skewed toward the funding of hardware acquisition and in-house analysts, rather that toward the leveraging of external centers of excellence who bear their own overhead costs. Worse, I do not see any positive steps to correct these deficiencies. Although the military intelligence community, and the scientific & technical intelligence analysts in particular, do well in exploiting open sources, the general approach of the intelligence community in recent years has been to create a duplicate architecture, to require analysts to go to separate workstations to use open sources, or to “bundle” detailed analysts queries into large “blobs” which force the private sector respondent to vomit back into the classified computers a mirror image of their entire holdings simply because they are not designed to “unbundle” the queries.

Developing OSCINT concepts and doctrine is critical to our future as an intelligence community. The National Foreign Intelligence Board (NFIB) was recently advised that 40% of the all-source product comes from OSCINT, at a cost of 1% of the National Foreign Intelligence Program (NFIP). While I personally feel that they underestimated the cost and that it is closer to $2 billion a year, 7% of the NFIP, this is never-the-less a shocking and instructive insight into the role that OSCINT plays in our profession.

Following are my policy recommendations on OSCINT:

a) Trying to get analysts to move to a different workstation, and bundling analysts’ queries, is dumb, expensive, and counterproductive. We must use our industrial partners, our traditional external research partners each of whom has a Sensitive Compartmented Intelligence Facility (SCIF), to provide both the security air gap, and the “laundering” of our specific queries, so that individual analysts working from their classified workstations can send out a specific query and receive a specific response at their classified workstation. We cannot afford to recapitalize the existing architecture, nor can we count on analysts to tolerate a diversity of workstations and the security requirements inherent in transferring OSCINT data from one system to another. Our industrial partners can be relied upon to provide the virus checking and other security measures associated with introducing external OSCINT into the Top Secret/Codeword domain.

b) Reliance on industrial partner outreach efforts is not enough. A national program is required which nurtures distributed centers of excellence in various subject matter areas. One of our most important deficiencies as a community is the inexperience of our analysts. The Director of Intelligence for CIA recently commented on his dismay, in assuming his position, at discovering that he had a bunch of “19 year olds” doing analysis. Obviously they are a few years older than that, but his point is well-taken–I can testify, as a result of my brief experience as an observer of the DCI’s Management Advisory Group, that the CIA strategy for keeping personnel costs down relies heavily on a constant cycling in and out of those under the five year mark. The result of this is a gap in the middle. In the military we match this lack of continuity and expertise by assigning people to different tasks every two years, never allowing anyone to master a language or a topic. There are two solutions: on the one hand, we need to develop a cadre of experts (and a related career progression path) in the many areas of interest; and on the other hand, we need to use our experts to oversee and take advantage of a much broader circle of distributed experts–experts whose overhead we need not finance. It is my opinion, expressed in draft legislation now in staffing, that the best way to nurture distributed centers of excellence is through a Vice Presidential initiative, a National Information Foundation, which provides grants to specific business, media, and academic institutions, and which is independent of, but responsive to, the intelligence community.

c) Finally, we need to extend our methods to the rest of government. It was instructive to me, during one Open Source Lunch Club (second Tuesday of every month at the Tivoli in Rosslyn) to hear a Special Assistant to the President comment that the intelligence community was fairly useless in strategic economic decision-making because it did not report on domestic economic matters, and that the rest of government, for instance the Federal Reserve, was fairly useless because it was lousy at the business of intelligence–of collection management and distilled analysis. We must overcome the existing separation between our intelligence community analysts with clearances, and the many expert analysts in different elements of government without clearances. One way to do this is to adopt Andrew Shepard’s recommendations, in his “Intelligence Analysis in the Year 2002: A Concept of Operations”, and put intelligence community analysts alongside the consumers. There are two reasons for doing this: the first is that our consumers need real-time decision-support, and not the flood of “classified information” that they now receive and ignore; and the second is that 90% of what a policy-maker reads and listens to today is both unclassified and unanalyzed. We have abdicated our role of informing policy, and by focusing on secrets, failed to exercise our critical capability to validate and place in context non-secrets.

Now let me turn to imagery and signals intelligence, before concluding with observations on the discipline nearest my heart, clandestine human intelligence.

Imagery intelligence, in its time, was the “oasis” in the desert. However, over the years, someone has come along and watered the desert. It was no surprise to me to see the President of the United States overrule the DCI and direct that one meter resolution synoptic imagery be pushed into the private sector. There are three good reasons why this was necessary: the first is that our restrictions to date have amounted to nothing more than a U.S. taxpayer subsidy of Russian and French imagery sales. The second, more important, is that our existing classified imagery systems and their related priority management process have proven unable to satisfy urgent contingency mapping, charting, and geodesy (MC&G) needs. It is my estimate that we are 75 satellite years behind in fulfilling basic MC&G requirements. In the case of the Marine Corps, of 69 countries of interest, there are zero 1:50,000 maps for 22 of our most likely contingencies, and dated (10-15 year old) maps for ports and capital cities only (not maneuver areas) for an additional 37 countries. Commercial synoptic (i.e. we can do contour lines) one meter resolution imagery is more likely to satisfy Marine Corps needs than the present government-owned constellation with its related priority system. Lastly, and also important, the reality is that we cannot maintain our technological leadership, our ability to produce satellites, under the reduced ordering rate now funded by the U.S. government–unless we are allowed to sell this technology overseas, we simply will not be able to keep the assembly lines open, nor continue to support leading edge research.

We still have not solved the wide-area imagery problem, although progress has been made. We need to be aware that Ted Turner, of CNN fame, is hell-bent of having his own imagery system and never again being censored by the U.S. government. It is inevitably that there will ultimately be a global constellation of commercial satellites with images for sale to anyone. This would be a good time to shift our emphasis from collection to processing, from stove-pipe analysis to integrated analysis and inter-disciplinary tipoff, and near-real-time tactical dissemination. That will provide our competitive advantage.

Signals intelligence (SIGINT) has always been a mystery to me. It was absolutely useless to me [as a case officer], and this was especially galling when a test run in Panama, done as a personal favor when I was chasing terrorists, demonstrated that SIGINT could make a dramatic contribution to my local operations. I simply was not important enough in the over-all scheme of things. We have a long way to go before SIGINT is a major contributor to local human intelligence (HUMINT) operations.

On a broader scale, the Clipper Chip initiative must be mentioned. There is no question in my mind but that NSA authorities–smart people with super computers–believe this to be essential to our future. However, and I say this as a direct result of my international notoriety and expanded circle of contacts, they are out of touch with the American people and with reality. The Clipper Chip is dead. NSA needs to focus on being able to do its mission secretly, without recourse to public admissions of incapacity–it also needs to consider how its smart, computer-friendly analysts with language skills, can make a difference in monitoring the unclassified traffic that is so prevalent on the Internet and other elements of “cyberspace”. NSA has a very important and expanded role to play in collecting and processing intelligence from cyberspace, provided we can all agree that much of the intelligence to be gained from cyberspace is not necessarily encrypted, but simply lost in the confusion and noise that characterizes that environment.

[And now to HUMINT, I will be blunt, for I care deeply about this discipline. When I was at The Farm they emphasized security and counter-surveillance. Then I got overseas and discovered that a) I was working from an official installation easily covered by host country surveillance; b) I was expected to work after hours and on week-ends, making it very easy for the host-country to identify me; and c) I was expected to meet every one of my 20 agents and every one of my fifteen developmental every week. There is no discipline more important to national security decision-making than HUMINT, for it is only “in your face HUMINT” from deep and proven penetrations which can persuade policy-makers to use force and take other measures which might be politically troublesome at home.

It is impossible to maintain a clandestine service worthy of the same from within official installations and using youngsters whose idea of cover is impossible to sustain because they have never really held a job in the private sector. Unfortunately, the existing funding profile for clandestine operations relies on official cover and a very young and inexperienced–and therefore low-cost–population. Going deep and long will cost money. It is my view that the funding profile for the clandestine service must double over time, with a majority of that doubling going toward increased non-official capabilities, increased support to pre-emptive military operations, and increased penetrations of “rogue-government” and international criminal gang leadership circles. This is the greatest challenge facing the U.S. intelligence community, and the one most likely to be ignored and under-funded.]

We have got to get out of the Embassies, and we have got to get away from the idea that only yuppies who can pass security hurdles should be allowed to be [case officers]. There is absolutely no justification for imposing the same security standards on collectors as are applies to all-source analysts. In my view, we need many more Third World ethnic career agents, many more [non-official cover officers], and much more in the way of contingency coverage. In brief, we need to go deep and go long. If I were DCI, I would spin off the Directorate of Operations (of which one third would be devoted to military support) as a separate agency, and I would give them the funding necessary to support mid-career hires–let someone learn their cover at the expense of others, and then turn them into spies. I would also ensure that we shifted to multiple networks of non-official covers with at least a bare bones overlapping coverage of every country and every global criminal organization and tribe in the world.

There is no question in my mind, but that in an age of technical sophistication, and national-cultural fragmentation, that HUMINT is move vital that ever. I have not touched on counterintelligence and the Bob Ames case. Nor do I need to. We fell in love with the polygraph, even though we knew in the mid-l980’s that it was a marginal vehicle. Two of my classmates got wrapped up in Cuba (and appeared on international television) because the Cuban double-agents all managed to pass the polygraph. There is a discernible tendency in the USA to rely too much on technical panaceas, and to eschew labor-intensive or “complicated” approaches. I am here to say: open sources are an important and fundamental part of the solution. Do NOT send as spy where a schoolboy can go. HOWEVER, if you do NOT have a diverse collection of spies out there, penetrating the wide variety of communal (non-national and often tribal or religious) groups that are hostile to your interests, then plan for major damage to national security and national competitiveness, for you have abdicated your right to survive and betrayed the most fundamental interests of the American people today and of the generations to follow.

NOTE: For detailed critiques of current U.S intelligence capabilities, and related recommendations for improvement, see among other publications the following:

“Intelligence in the 1990’s: Recasting National Security in A Changing World”, American Intelligence Journal (Summer/Fall 1990)

“Applying the ‘New Paradigm’: How to Avoid Strategic Intelligence Failures in the Future”, AIJ (Autumn 1991)

“The National Security Act of 1992”, AIJ (Winter/Spring 1992)

“A Critical Evaluation of U.S. National Intelligence Capabilities”, International Journal of Intelligence and Counterintelligence (Summer 1993)

“Talking Points for the Director of Central Intelligence” dated 20 July 1993, subsequently published in Proceedings of the Second International Symposium on “National Security & National Competitiveness: Open Source Solutions” (Washington, D.C. 2-4 November 1993)

“Reinventing Intelligence: Holy Grail or Mission Impossible”, Periscope (Journal of the Association of Former Intelligence Officers), June 1994