CIA ANALYSTS AND NSA

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP79M00467A002400030012-0
• Release Decision: RIPPUB
• Original Classification: T
• Document Page Count: 4
• Document Creation Date: December 16, 2016
• Document Release Date: December 3, 2004
• Sequence Number: 12
• Publication Date: August 12, 1976
• Content Type: MF

E
ROUTI
TO
: NAME AND ADDRESS
DATE
INITIALS
f curve-
4s~
2
3
4
ACTION
DIRECT REPLY
PREPAR
E REPLY
APPROVAL
DISPATCH
RECOM
MENDATION
COMMENT
FILE
RETUR
N
CONCURRENCE
INFORMATION
SIGNATURE
REMARKS:
FROM: NAME, ADDRESS, AND PHONE NO.
DATE
'21)11
1
Top Secret
(Security Classification)
COMINT
oved W
CONTROL
Access to this document will be restricted to
those approved for the following specific activities:
Warning Notice
Sensitive Intelligence Sources and Methods Involved
NATIONAL SECURITY INFORMATION
Unauthorized Disclosure Subject to Criminal Sanctions
Top Secret
(Security Classification)
Approved For Release 2004/12/20 : CIA-RDP79M00467A0024000300b~1-3hIPDET
 Approved FWelease 2004/BETA-RDP79M00.A002400030012-0
IV 4 cr) :+
12 August 1976
-MEMORANDUM FOR: Dr. Sayre Stevens
Deputy Director for Intelligence
Chairman, SIGINT Task Force
CIA Analysts and NSA
25
1. I spent last week taking a preliminary look, visiting with OSI, OWI,
OSR, OER, OPR and OCI. I thought you might be interested in the preliminary
findings. They are not unexpected.
2. Almost everyone complains, with varying degrees of emphasis, that:
a. NSA gists are not complete enough and someti :yes miss
the more important points to be found in the raw transcript;
b. NSA analysis is incomplete, unreferenced and often
indiscriminantly mixes SIGINT. and collateral;
c. NSA does not like to put out gists or analyses until the
NSA analyst is satisfied with the completeness of his data or
until he understands the significance of the SIGINT he is gisting.
In some cases, this produces delays of sixteen months;
251
Approved For Release 2004/CA-RDP79M00467 0409p011~ 25!
 Approved F elease 2004/12/ DP79MOON A002400030012-0
Approved For Release 2004/1 rCt RDP79M00467A002400030012-0
3i. i
would like more data but finds NSA analysis generally useful and welcomes
its tendency to aggregate data. Nor is timeliness much of a problem to OER.
I judge this to be a reflection of OER's greater use of collateral, preference
for aggregated data and different emphasis on analytic problems. OCI says
that it enjoys a generally excellent relationship with NSA, has no trouble
getting the full text of messages and finds ,the timeliness, of NSA reporting
good. It does not deal much with military affairs and experiences few of
the frustrations expressed by OSR or OWI. This may be because of NSA's
interest and better understanding of current intelligence.
4. Asked about the usefulness of stationing an NSA liaison team at
CIA. almost everyone agreed this would be helpful in the highly compart-
mented areas, although some analysts expressed concern whether this would
still separate them from data they would like in their own files. They did
not find so much utility in the more general COMINT area. The research
d. NSA quite often refuses to provide transcripts or
"technical" data. When the CIA analyst does get this material,
he often most do it "via the back door, " relying on his friend-
ship with his counterpart at Fort Meade; and
e. "Back door" information cannot be used in CIA publi-
cations except through subterfuge. The information is obtained
In advance; the analyst must wait for NSA to publish it. This
can Introduce significant delay.
3..... While the vehemence of complaints vary, five of the nine groups I
..talked to were pretty damned excited. The two major exceptions were OER
and OCI, although OPR China analysts are reasonably well satisfied. OER _
offices, except for OSR, were enthusiastic about stationing a CIA exploitation
team at NSA to carry out the same function there as does IAS at NPIC. OCI
itself saw no need for this. OER was also cool since it feels satisfied with
the way things are. Everyone was concerned that institutional changes might
disturb established analyst-to-analyst relationships.
5. Analysts have not thought too much about the audience for which they
are writing. Most agreed that there is a difference between production aimed
at the policy maker and production aimed at their production counterparts.
When the subject was brought up, most agreed that they often put too much
source information Into reports for policy makers, the very ones who were
most likely to compromise these sources. No one disagreed with allowing
NSA to help them reduce this danger, a side benefit of having an NSA team
here. No one used the word "censorship."
 Approved FRelease 2004/1 I2b: CIA-RDP79MOV A002400030012-0
ffy
25
6. All of the above judgements are qualitative although numerous and
concrete examples of CIA-NSA problems were cited. The Task Force plans
to return to the issue later on in Its work.
25X1
1A
1A-
Q CIA Analysts and NSA
., 25
Original - DDI
I - DDS&T
1 - Comptroller
Approved For Release 2004/'gRTA-RDP79M00467A002400030012-0