TRANSCRIPT OF DDI REMARKS TO THE SECURITY AFFAIRS SUPPORT ASSOCIATION

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP95M00249R000801140011-6
• Release Decision: RIFPUB
• Original Classification: K
• Document Page Count: 15
• Document Creation Date: December 22, 2016
• Document Release Date: March 27, 2012
• Sequence Number: 11
• Publication Date: November 13, 1984
• Content Type: MISC

Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
TRANSCRIPT OF
DDI REMARKS TO THE
SECURITY AFFAIRS SUPPORT ASSOCIATION
13 NOVEMBER 1984
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
I'd like to thank you very much for the invitation to stand in
for John McMahon. I bring you greetings from Bill Casey, the
Director of Central Intelligence, and from John McMahon as well. I
could be secretive about why John McMahon couldn't come, but un-
fortunately, thanks to the newspapers, you know everything we are
doing--and a great deal we are not.
I'm very glad to be here with you tonight. I must say that
after the drive out from San Francisco this evening, I'm glad to be
anywhere tonight.
Speaking to you on software is particularly inappropriate for
me, as a specialist in Russian studies and an historian. But I am
no stranger to inappropriate things. When I was on the NSC staff
during one administration that shall remain nameless, I was present
when the President of Italy attended a state dinner in his honor.
The entire White House was decorated, inappropriately, with
thousands of yellow chrysanthemums--the Italian flower of death. I
was present when, inappropriately, at a state dinner for German
Chancellor Schmidt--a visit where the principal topic of con-
versation had been offset for our troops in Europe--the afterdinner
entertainment was Joel Grey singing highlights from Cabaret, ending
with the song "Money, Money, Money."
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
So even though my talking to you about software is inap-
propriate and potentially embarrassing, to use the old saw, my job
here is to speak and yours is to listen and with any luck we'll
finish at about the same time.
This past summer, the Security Affairs Support Association
became a professional organization, a change that allows government
employees to seek membership. This confirms the goal for which this
organization has been well known, to create an organization in which
the security affairs interests of government and industry can find
common meeting ground to discuss issues of mutual concern. This
arrangement makes good sense, because it takes advantage of the
talents, capabilities, and initiatives in both the private sector
and in government. We all know that neither has monopoly on brains
or talent and certainly not on money.
This country is unique in the ways in which government and
private industry work together. In our work we have long depended
on you and the companies you represent and many others to build
large technical collection systems and design and build computer
hardware and software. Beyond this, however, we also have depended
on your expertise and your facilities to help us understand and
forecast Soviet weapons systems. More recently we have been turning
ever more frequently to the private sector for assistance on a broad
range of other areas, from new technologies to international
economic problems to scores of other subjects covering more than
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
fifty different disciplines. Just this year, more than 1,200 CIA
analysts attended more than 500 conferences, many of which were
sponsored by your companies and others in the private sector.
In the U.S., we rely on a combination of patriotism and profit
motive to make our system work. And I believe the evidence suggests
that the system serves us well. New technological devices and new
analytical techniques that enable us to understand growing threats
to U.S. and its people are based on the synergistic nature of the
relations between private industry and government.
I could cite several examples to support my contention that our
system of private and public developments work well. Our achieve-
ments in space certainly prove the point--and our ability to create
new and more effective defense systems also requires close working
relations between the public and the private sector. There is
little question that developments in computer world count as well.
Before addressing the major subject of your conference, and of
my talk on Agency software, I would like to give you briefly a more
general picture of where we've been and where we're going. The last
four years have seen remarkable growth in the Intelligence
Community's budget. Now while this now leveling off, it has enabled
us to restore many of the capabilities that we lost in 1970s. These
funds have allowed us to bring on board operations officers,
analysts, technical specialists, and support personnel.
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
Speaking very narrowly of my own organization, the Directorate
of Intelligence, our reorganization three years ago--undertaken, by
the way, by John McMahon, my predecessor--to integrate our analytic
disciplines has created a quantum jump in both the quantity and the
quality of analytic products we can deliver to policymakers. Our
analysis has taken on a new dimension because political, economic,
and military analysts--and analysts from other disciplines as
well--are working together to create a multifaceted approach to
problems.
In the last four years, we in CIA for the first time have
developed and implemented a comprehensive research program covering
a staggering number of countries and issues. We devote now about
half our analytic resources to the Soviet Union, from its economic
and political problems to its military strength, future weapons
programs and strategic intents. The other half of our effort is
focused on problems as diverse as terrorism, narcotics, political
instability in key countries, international energy and resource
development, nuclear nonproliferation, the gray arms market,
technology transfer, forecasting food supplies, tracking wars and
insurgencies worldwide, monitoring Soviet compliance with arms
control agreements, scientific and technical developments worldwide,
and many more. We produced more than 800 major research assessments
last year, along with our current intelligence. More importantly,
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
this dedicated research effort is establishing once again the strong
base of information, data, analysis, and expertise to support the
policy process.
The production of national estimates and the speed with which
we are producing them has also increased. Estimates can sometimes
be the source of controversy. You can read about most of them in
the newspapers, but we are making an effort to provide decision
makers with the message they have to have--not necessarily the
message they want to hear. I can report to you that in recent
years, thanks primarily to the efforts of Bill Casey, a spirit of
real cooperation and mutual respect perhaps unique in the history of
the Intelligence Community has developed among the principals of the
intelligence agencies. One reason for this has been a strong effort
to insure that all hypotheses, alternative scenarios, and con-
clusions are heard and reported to the policymakers. Those agencies
that disagree with the majority view spell out their differences in
the estimate. And we no longer relegate those minority views to
footnotesn but they are embodied in the full text of the
estimate. We make tough calls, calls which are sometimes painful
for policymakers. But we are meeting our commitment to provide the
best judgment without partisan flavor or political taint.
On the collection side, progress also has been made. We have
rebuilt our human intelligence capability after the severe cuts of
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
the last administration. While the young people we are hiring today
lack some of the language capability and area knowledge we would
like, still they are remarkably sharp and demonstrate a talent and a
drive that gives us confidence in our future capabilities. Indeed,
I would say that sometimes their skills are awesome.
CIA received more than 150,000 inquiries about employment last
year. Our recruiters are meeting receptive audiences on college
campuses these days for the most part. And it is interesting to
note that there appears to be renewed interest in government service
in a more general way.
On the technical side, investments in new technical systems are
beginning to pay off. This means a tremendous increase in the
quality and precision of intelligence information we can collect.
It also means that the volume of data will substantially increase.
This suggests that we must begin to invest in processing systems to
match our capability in collecting raw data.
In many ways, in my view, we are better prepared to process and
use information from existing and new technical systems than we are
to deal with the growing volume of human source reporting, and most
daunting of all, the tidal wave of increasingly vital overtly-
available information, particularly in the scientific and technical
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
arenas. We have to avoid being overwhelmed by this volume of
information from all sources, and that is where developments in
computers, software, and other new technologies really must help us.
As I indicated at the outset, this is a very tough audience for
me, student of Russian history, to address on software, since most
of you are experienced professionals and are well versed not only in
sophisticated computer techniques, but you know a good deal about
intelligence applications as well. But like many managers in
industry and government, I am learning quickly. I'd like to make a
few brief remarks about the present state of our computer activity
and where we are headed.
First, and I suppose, partly as an element of self confession,
let me amplify just for a moment on a problem I mentioned lightly a
moment ago. A serious problem for both American industry and
government is the computer illiterate senior manager. To tell the
truth, and I don't think I'm telling you new truth, important
decisions on computers and ADP equipment are made by managers who
hardly know a mainframe from a Mack truck. What do we do? We turn
to the computer specialists in our own organizations who think
narrowly and protect turf and for whom larger scale planning, net-
working, and experimenting is anathema. And so we sometimes develop
inadequate systems that can't talk to each other and meet only
today's needs. It seems to me that an imperative course, not just
for those of us in the Intelligence Community, but in industry
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
as well--and your group may be the one exception--is to bring senior
managers into the computer age and enable them to know enough to use
effectively experts and then to make sensible, broad-ranging,
future-oriented decisions and guidance.
Now let me turn to CIA specifically. The major problem in the
use of computers at CIA revolve around compartmentation and
security. Unlike organizations of similar size in the private
sector, we have to have a system that operates on a need-to-know
basis, and that may involve only a handful of people. We must
protect ourselves at the same time against "hackers" from the out-
side, and the possibility of "moles" from the inside. This is a
major reason why we have spent so much time and money developing our
own software--we have found that many private sector applications
are not sufficiently stringent or secure for our purposes. We also
realize that, from the perspective of private industry, the costs of
developing a package for government that has no other application
often is not profitable.
The CIA is now operating five major computer systems, all using
the same general architecture, but each almost completely in-
dependent of the other. We have a system for analysts, one for
operations, one for administration, one for physical and personnel
security, and one for processing and analyzing data from technical
collection. Each of our systems use similar but distinct software
applications, all developed in house. For example, our admini-
strative system includes electronic mail. We had to develop
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
software that would allow access based on individual clearance and
need to know: the system had to be impervious to serious efforts at
penetration as well as idle curiosity.
In contrast, though, some of our needs not so unique. The
agency is moving away from reliance on huge mainframe systems to
increasing use of computer networks and even self-contained personal
computers. We envision systems where many professionals will rely
on personal computers for much manipulation of data, and where
standard forms of software--spread sheets or word processing
programs--can be bought over the counter. This will allow us to
shift from software designed by ADP professionals for other, if you
will forgive me, "computer freaks" to systems that are more user
friendly. Indeed, the use of personal computers may create new
security problems for us--we may have to guard not only against
analysts taking their papers home with them but also their floppy
disks.
We have high hopes that the fifth generation of computers--and
developments in Artificial Intelligence--will have great application
at CIA. We will rely on A.I. in expert systems applications to
enable us to detect indicator anomalies for warning, to synthesize
combinations of data for analysis, to scan mail to pick out critical
messages, or to pick out gaps in our knowledge. Applications of
A.I. in processing huge quantities of raw data
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
without having to translate raw data into standardized formats, as
we now do should help separate the wheat from the chaff, especially
in SIGINT and imagery.
A.I. should do some things for us that will enhance our ability
to support the policymaker. More sophisticated simulation and
modeling techniques will increase our ability to predict alternative
outcomes of future events. A.I. should help analysts compare dis-
similar forms of data--imagery, SIGINT, regular text--without having
to put everything in a standard format. A.I. might help us package
our inputs to the policy process in more usable forms to meet con-
sumer needs, and A.I. may help shorten the production cycle.
Another application might involve accessing more data on a real time
basis, especially in crises. So, it appears that we are only
beginning to tap the potential of the next computer generation.
Frankly, progress in the area of A.I., in our view, is likely
to be painfully slow. Promises of quick advances with practical
applications should be treated with some skepticism. We understand
the reasoning process that experts use to generate judgments, but
much is "intuitional," especially when dealing with softer data. I
can't tell you how many contractors have tried to sell me software
that will enable me flawlessly to predict the next action of the
Soviet leadership. Until we can understand more about how the
intuitional process works, it seems to me, it will be difficult to
write "expert systems software" that can duplicate what
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
analysts do. Despite these problems, A.I. offers much promise, and
we will look to the research and industrial communities for advances
in these areas.
On the operations side, we hope to move toward a paper-free or
almost paper-free environment. This is especially important in
protecting overseas systems, although I might add that judging from
the newspapers, it would leave a number of Iranians unemployed who
together the papers that went through the
are still trying to4ttp
shredder in 1979. A similar system on the production side may
enable us to develop a high-quality product for key consumers
without going through the printing process. And frankly, I believe
this outcome is nearer than virtually any policymaker believes at
this point.
CIA, like many private firms our size, is suffering a great
software applications backlog. Just as in the private sector, we
have too few programmers and too many jobs. Our present back-log
probably amounts to perhaps as much as some 400 man-years.
Clearly, we should be working together to develop more
effective and efficient programming methods. One way to solve the
problem involves more end-user programming, and this will become
more practical as we move away from large central systems to desk--
top machinery. We will also be aided in the future by the fact that
a new generation of employees is increasingly computer
11
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
literate. New professionals are running about 50 percent computer
literacy, thanks I suppose to video games and the presence of
computers in high schools and colleges. We've even gotten our most
senior officers to begin using computers--we've all had to take some
training and learn some simple computer language so that we can
communicate with each other without having to meet in Executive
Dining Room.
In terms of the Intelligence Community, we have developed
several ways of working together, although much of our software has
been developed on an internal agency basis. The Community does try
to share ideas, information, and technology, including software
where possible, although we do have to maintain compartmentation.
Air Force and DIA have developed a modular architecture called MAXI
for sharing intelligence and processing messages for analysts.
Eventually, the Community will have some 33 facilities for such
sharing--15 of these are now on line. Shared SAFE systems between
CIA and DIA are now completing Phase I which permits shared message
retrieval capability. Phase II will take us into 1987-88 time frame.
The Intelligence Community is also sponsoring the Community
Information Retrieval System. If successful, this will bring
together processing assets from five networks at DIA, NSA, Air
Force, CIA, and NPIC. The major difficulty involves security and
standardization problems. The Community is also working toward
developments in Artificial Intelligence. CIA's Dr. Phil Eckman
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
chairs the Artificial Intelligence Steering Group--with membership
from a dozen agencies and federal components. Because of costs
associated with A.I., we want to avoid duplication of effort and
circulate information about the subject. CIA will sponsor its third
annual Community-wide symposium in March of 1985. This will bring
together members of the Intelligence Community and private industry
to exchange ideas of mutual benefit.
ADP now consumes a substantial portion of CIA's resources and
space. We're now well along in construction of a new headquarters
facility to accommodate people displaced by computers. Our ex-
perience is probably typical of organizations our size, and our aim
is to stabilize ADP costs and perhaps even drive them down One way
to do this is to push the state of art in selected critical areas
and not wait for it to catch up with our needs. This brings me back
to theme I raised at the beginning of my remarks. The synergistic
nature of relations between intelligence organizations and private
industry should boost the speed--although I hope not the cost--of
solving software problems.
Let me conclude on a broader note about why all of this
matters. I believe strongly that the Intelligence Community
increasingly is the only place in the national security arena that
is devoting requisite energy and resources to examining the longer
range problems that will face our country in 5 or 10 or 20 years.
Indeed, with our unique and comprehensive collection and analytical
13
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
 Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6
capabilities we may be the only element of government (and perhaps
industry too) able to cross the boundaries of bureaucracy, sub-
stance, and specialization to give our leaders the information they
need to guide the country in an increasingly interrelated and
complicated world.
Whether we are talking about future Soviet weapons, foreign
energy supplies, the impact of new technological developments oil
Third World exporters of. old raw materials, great new immigration
flows around the world, nuclear proliferation, famine in Africa, or
the impact of new foreign developed technologies on our national
security, we are in the forefront, helping to inform, warn, and
educate our leaders. But information is key. We cannot do this
without great advances in our use of computers for information
handling, new analytical techniques and, ultimately, A.I..And that
means software; and that means you. We need you. We welcome your
partnership in service to the nation.
14
Declassified and Approved For Release 2012/03/28: CIA-RDP95M00249R000801140011-6