<<

. 82
( 118 .)



>>

electronic filter large volumes of printed matter, including transcripts of



465
speech and data from Internet discussion groups. One of the sample
questions in the test was "What have the effects of the UN sanctions
against Iraq been on the Iraqi people, the Iraqi economy, or world oil
prices?" Initial tests proved very successful, increasing the ability to
locate target information from 19 percent to 27 percent in just one year.
Far more difficult than machine translation of printed texts is
automatic translation and transcription of voice communications, such
as intercepted telephone conversations in a variety of languages and
accents. The ability to automatically spot targeted words in millions of
telephone calls all over the world has long been a goal of NSA. A recent
breakthrough was made by biomedical engineers at the University of
Southern California, who claim to have created the first machine system
that can recognize spoken words better than humans can. The research
was largely funded by the Pentagon, long used as a cover for NSA
contracts.
According to the university, the system can "instantly produce clean
transcripts of conversations, identifying each of the speakers." Known at
NSA as "Speaker ID," the USC's Berger-Liaw Neural Network Speaker
Independent Speech Recognition System can mimic the way brains
process information. This gives the computer the ability to conduct "word
spotting" in target communications regardless of who or what
pronounces the word.
The new system is also far better than the human ear at picking out
words from vast amounts of white noise. It can even extract targeted
words or conversations from the background clutter of other voices, such
as the hubbub heard during conference calls, meetings, or cocktail
parties. "The system can identify different speakers of the same word
with superhuman acuity," said university officials.
Despite such progress, by 2001 there was still far more traffic than
there were people or machines to handle it. "It's a good-size problem,"
said Hayden. "It's one that we're paying attention to, but the fixes are not
immediate. There's probably no philosophers' stone here that we can
touch and say, 'Oh yes, now the linguist problem's fixed.' There's
probably a whole bunch of discrete decisions that you make that you
begin to reduce the magnitude of the problem. One aspect of the problem
is, just given the nature of our business, the demands on linguists are
higher."
Some NSA language training takes enormous amounts of time, said
Hayden, who himself was trained as a Bulgarian linguist. "Group Three
languages, and I believe that's Arabic and Hebrew, take eighteen
months," he said. "And Group Four languages take two years. And those
are Chinese, Japanese . . . And then there is a whole other addition there
to turn someone who has working knowledge of the popular language



466
into a cryptolinguist, which is the specialized vocabulary. . . . It's a long
time, these are long-term investments. And you can see why, then, we
have trouble mostly with our military linguists who move a lot, whereas a
civilian you hire for thirty-five years and you make a front-end
investment of five years, you've still got thirty years of return. You've got
a GI going through here on an eighteen-month tour."
Realizing NSA's personnel plight, the House Intelligence Committee
began a major push in the late 1990s to redirect money away from
various fields, ranging from satellites to support staff, and toward
analysis and linguists.
"We need to hire a lot more people than we have authorized strength
to do," Terry Thompson told a group of employees in late 1999. "The DO
has recently told the Human Resources Review Group that they would
like to hire twenty-six hundred more people to do language work and IA
work, Intelligence Analyst work. And the reason for that is, if they look at
their attrition projections, they expect to lose about a thousand people
over the next couple of years and so they want to hire those people back.
And then they want a plus-up of about sixteen hundred people over and
above that, just to be able to do the work that comes in today." According
to one senior NSA official, the agency hired about 698 people in 2000.
For 2001, Congress gave NSA an additional $3 million to go toward
hiring, plus $3.5 million more to use for signing bonuses for particularly
desirable candidates.
Just as the fall of the Soviet Union created a need for exotic
languages, the proliferation of low-cost, complex encryption systems and
fast computers has forced NSA to search for more mathematicians whom
they can convert to codebreakers. In a series of lectures at NSA in the
late 1950s, William F. Friedman, the father of modern cryptology, argued
that cryptology should be considered a separate and distinct branch of
mathematics. It is little wonder, therefore, that NSA employs more math
majors than any other place in the country, and possibly the world.
Thus the national decline in math test scores, the decreasing focus on
math in the classroom, and the paltry number of graduate students
seeking doctorates in the subject have become major concerns within
NSA. "The philosophy here is that unless the U.S. mathematics
community is strong, healthy, vibrant," said James R. Schatz, chief of
NSA's mathematics research division, "then we're not going to have the
kind of population to recruit from that we need."
Some at NSA trace the growing scarcity of mathematicians back to the
early 1980s. It was then, according to one agency official, that "the
agency succumbed, as did the rest of the American society, to the
increasing gap between its population of technical specialists and a
generalist population." As the last editor of the NSA Technical Journal,



467
which ceased publication in 1980, the official witnessed the decline in
mathematical and scientific education firsthand. It was one of the
reasons for the Journal's termination, he said, noting that many of the
contributions were becoming increasingly "irrelevant to (and
unintelligible to) all but a small audience." He added that if Friedman
was correct in including cryptology as a branch of mathematics, "then
large numbers of NSA's employees, even at the professional level (and
within the professions, even within senior positions), are ill-equipped for
their trade."
In an effort to reverse the trend, NSA recently launched a new
program to seed the academic soil in order to keep the supply of
mathematicians coming. It involved providing $3 million a year, through
research grants, to mathematicians and also to summer programs for
undergraduates. Yearlong sabbaticals at the agency were even offered to
promising number lovers. In a rare foray into the unclassified world,
then-director Minihan expressed his worry to a convention of
mathematicians in 1998. "The Cold War is characterized by battles not
fought, lives not lost," he said. "That era was fought with mathematicians
arid cryptologists."
"Over a three-year period," said Schatz optimistically in 1998, "we're
going to be hiring over a hundred mathematicians with Ph.D.'s. There's
nothing like that in the world, really. A university might have one or two
openings a year, if that." But just as NSA seems to be getting its need for
mathematicians under control, it is facing an even more daunting task in
recruiting enough computer scientists. Among the problems, according
to Michael J. Jacobs, chief of NSA's codemakers, is 42 percent fewer
graduates with computer science degrees now than in 1986.
Among the most sensitive issues facing NSA in the post”Cold War
period has been the hiring, as well as promotion, of minorities and
women. For years NSA has had serious problems keeping up with the
rest of government”and the rest of the intelligence community”in such
employment statistics. "I have been here at NSA for over twenty years,"
wrote one frustrated employee in the mid-1990s, "and as a minority,
have experienced racial discrimination like I have never seen before. The
minorities here at NSA are so very stigmatized by the 'Do nothing,
powerless' EEO [Equal Employment Office] and the IG [Inspector
General] organizations . . . there is no adequate or effective process for
minority complaints here at NSA. Many racial discrimination and fraud
cases have been reported/presented to NSA's EEO and IG, and nothing,
absolutely nothing, has been done."
Another complained, "EEO is a joke. . . . Nothing is held confidentially
or anonymously. Retaliation is common and well known around the
Agency. Most African Americans have stopped complaining and warn
younger, less experienced African Americans against complaining in fear


468
of retaliation and retribution." And still another cautioned, "It is a well
known fact that if you stand up for your rights it can be a crippling
experience, but become a whistle blower, and your career will experience
the Kiss of Death!"
In a 1988 study of the intelligence community, done at the request of
Congress, the National Academy of Public Administration found women
and minorities underrepresented at NSA. Two years later, the Senior
Advisory Group, a group of senior black NSA employees, examined the
barriers faced by African American applicants and employees in hiring,
promotion, and career development. They gave the agency low marks,
citing institutional and attitudinal barriers. And in 1993 the Equal
Employment Opportunity Commission concluded that little had been
done to correct problems identified five years earlier. Finally, in 1994,
both Congress and the Pentagon's inspector general hauled the director
in for questioning as to progress in hiring and promoting minorities and
women.
A key problem, the Department of Defense inspector general pointed
out, was the tendency of NSA recruiters to go after the "best and the
brightest." "The philosophy," said one senior personnel manager, "is that
it is better to hire an applicant with a 3.2 grade point average from
Stanford than one with a 4.0 from a school you've never heard of."
Although the former strategy keeps the agency well endowed
intellectually, it does not help the agency correct its racial and gender
imbalance, it was argued.
NSA did make some efforts to recruit minorities, but more often than
not they were only halfhearted. In an effort to recruit Hispanic students,
the agency set up a Southwestern Recruiting Office in Phoenix in 1989.
However, instead of staffing it with a Hispanic recruiter, the agency sent
a sixty-year-old black male. The result was a total of eleven people hired
in three years”none of whom were Hispanic. The office was closed in
1992.
For Director McConnell, the problem lay in the numbers. Although in
1993 women made up 43.4 percent of the federal workforce, at NSA they
represented only 36 percent. And while 27.7 percent of federal
government employees were members of minority groups, NSA's minority
representation stood at a dismal 11 percent. In his agency's defense,
McConnell pointed to the highly technical nature of its work”
mathematics, engineering, computer science, and language: "skill areas,"
he said, "in which minorities have been traditionally underrepresented."
For example, McConnell noted, "we have probably the highest
concentration of mathematicians in the country." But "of the 430
doctoral degrees in mathematics awarded to U.S. citizens in 1992, only
11, or 2.5 percent, went to minorities," he said. "Can you imagine the



469
competition for that 2.5 percent between companies like IBM or GM or
whatever and NSA? It's very, very stiff competition."
To help correct the imbalance, McConnell established a policy of
encouraging his recruiters to make one-third of their new hires
minorities. In fact, the recruiters exceeded the quota, achieving 38.3
percent minority hires. But with NSA hiring fewer than 200 full-time
staffers a year between 1992 and 1996, the quota system at this late
date amounted to little more than tokenism. In the meantime, McConnell
was left to deal with complaints from the agency's white males, who
make up 57.5 percent of the workforce. Although no "reverse
discrimination" lawsuits had yet been filed, McConnell was holding his
breath. "So far I haven't gone to court," he said. "Time will tell."
In an effort to ease tensions, an Office of Diversity Programs was
established to help ensure that minorities were fairly represented in
programs throughout the agency. Among the units of the office is the
Alaska/Native American Employment Program, which in 1999 sponsored
a presentation by storyteller Penny Gamble Williams, the tribal chief of
the Chappaquiddick Indian Nation of the Wampanoag Indian Nation,
relating tales passed down through the generations. A luncheon of
buffalo meat in the Canine Suite followed.


After more than four years in the director's chair, McConnell retired
on February 22, 1996. His replacement was Kenneth A. Minihan, a tall,
broad-shouldered Air Force lieutenant general. Unlike McConnell, who
had spent most of his career in staff (as opposed to command) positions,
General Minihan arrived at NSA after running two previous intelligence
organizations: the Air Intelligence Agency and, briefly, the Defense
Intelligence Agency. He was born in 1943, the same year as McConnell,
in Pampa, a dusty, oil-soaked town straddling the old Santa Fe Railroad
in the Texas Panhandle. After graduating from Florida State University in
1966, he entered the Air Force as an intelligence officer, serving in
Vietnam, Panama, and Italy and in a variety of positions in the Pentagon
and at Air Force Headquarters.
In 1981 Minihan went to NSA as chief of the Office of Support to
Military Operations and Plans. He also served in the agency's Directorate
of Operations, as commander of the Air Force's 6917 Electronic Security
Group. Minihan was named director of DIA in July 1995; there, one of
his chief assignments was to review tainted information about Russian
weapons systems passed by the CIA to the Pentagon. The Pentagon had
received this bad intelligence because of the massive compromise of
American spies in Russia by CIA turncoat Aldrich Ames.
According to Minihan, NSA's problems had become a great concern to
both Secretary of Defense William J. Perry and CIA director John


470
Deutsch. "They would use the phrase 'NSA doesn't get it,' " he said. "And
they were somewhat impressed with how I was beginning to take over the
reins of DIA, in the sense that we 'got it' at DIA." Thus the decision was
made to shift Minihan to NSA. During the transition to his new job,
Minihan spoke to a great many people both inside and outside
government about the agency and was stunned to find that the reaction
was virtually universal. "I would say I spent a good month or so talking
with lots of people," he said. "It was almost riveting in the common sense
that they all expressed that we [NSA] don't get it."
Once in place at Fort Meade, it didn't take long for Minihan to
understand why this was so. "It . . . really surprised me, both how
accurate Dr. Perry and Dr. Deutsch were . . . ," he said. "In my mind we
had fallen into a”I've never used this phrase before”sort of like a loser's
mentality, a loser's mind-set." One cause, said Minihan, was the
constant downsizing: "We'd lost about a third of our workforce. What we
had done is we were accepting the loss of program and people resources
as a norm. You've got another three percent cut. So we're going through
our tenth straight year of three percent decline. And we just accepted
that." Another early concern for Minihan was finding a new deputy.
When he arrived, the position was occupied by William P. Crowell,
appointed by Admiral McConnell two years earlier. A native of Louisiana
with an impish grin and a taste for Cajun shrimp, Crowell joined the
agency in 1962 and rose quickly, a decade later becoming chief of A
Group, the section responsible for attacking Soviet cipher systems.
Crowell foresaw the enormous impact that the personal computer would
have on both society and NSA and pushed the agency to begin taking
advantage of commercial, off-the-shelf technology. This was the key, he
believed, to improving both the way NSA attacked code problems and the
way it disseminated the results. Eventually rising to deputy director for
operations, Crowell championed the Intelink, the highly secret

<<

. 82
( 118 .)



>>