In defense of the beleaguered spysat
The difficulties of humint
Another problem with the claim that human intelligence has been neglected in favor of satellites is the fact that sometimes it is all but impossible to develop human assets inside a country or organization, no matter how hard the intelligence community tries to do so, and technical intelligence methods, particularly satellites, are the only option that can produce any intelligence at all. Journalists tend to be unsympathetic to this reality and point to a lack of human sources as a sign of intelligence failure at best, or extreme incompetence at worst.
During the 1960s no matter what the CIA did it could not get a human asset into China’s nuclear program. The Chinese nuclear program was bigger than al Qaeda, and the United States knew where it was primarily located, and yet U.S. intelligence agencies could not penetrate it; and unlike with the Soviet Union, there were no walk-in Chinese atomic spies. But was this an intelligence failure or simply the harsh reality of trying to penetrate an organization that is actively resisting such efforts and is ideologically and culturally (not to mention ethnically) extremely different from ourselves? While acknowledging that satellites have their limits, the critics often overlook the fact that there are limits to humint as well, and that the espionage business is nothing like James Bond movies.
As a result of its inability to get agents into the Chinese nuclear program, the CIA had to rely upon technical means because they provided some information, and some information was immensely better than none at all. So is the story of American intelligence collection about the Chinese nuclear program a story of the failure of humint or the success of technical intelligence?
The reality is that the CIA has developed substantial human assets throughout its history and doesn’t get the credit for this that the agency should get. The myth that the United States intelligence community is no good at human intelligence is fueled in part by the publicity that the capture of Soviet spies like Robert Hanssen or Aldrich Ames produces, and the relative lack of publicity that the capture of American spies, like Igor Polyakov, produces. This creates a misperception that the other guys (the Soviets, Chinese, Cubans, Israelis, etc.) are good at human intelligence whereas the United States is not. When Aldrich Ames was caught by the FBI, it was on the front page of the New York Times. It is not hard for critics to take the next step and allege that one of the reasons the intelligence community must be so bad at this is because it has gotten lazy and relies upon its satellites as a crutch.
In addition, the claim that the U.S. spends too much money on spy satellites and not enough on humans ignores the fact that the money requirements for the two are not only not similar, they are not even comparable. A satellite reportedly can cost up to a billion dollars or more (although that rounded dollar figure is probably an anachronism). Human intelligence assets are much cheaper, so buying more of them is not going to cost much money—assuming that they can be bought and that money is the decisive factor in their recruitment. Nor can spy satellites be funded with cutbacks in human intelligence; the amount of money that the CIA spends paying human intelligence sources around the world probably would not fund half a satellite, let alone a whole one. Satellites and human sources simply don’t coexist on the same budgetary plane and nobody should pretend that human sources are neglected due to the high costs of satellite spies.
As recent history has demonstrated, human sources have some major weaknesses of their own. They are limited by their location and perspective in the organizations that they inhabit—like the blind man feeling the tail of an elephant. And humans lie, they embellish, and they guess, requiring the intelligence analyst to gather even more information simply to validate the source.
Take for example the case of Anatoli Golitsyn, who in 1963 defected to the United States and provided not only useful information about Soviet spies within NATO, but also false information and bizarre theories that created considerable confusion within the CIA and the British spy agency, MI-6. Or take the case of the four sources who provided false information on Iraqi biological weapon production trailers. Although a satellite photo might tell you a limited amount of information, nobody doubts its honesty, and nobody believes it has an agenda.
Finally, what this argument neglects is the fact that the United States is very good with technical collection and nobody else in the world even comes close. The true imbalance is not American satellites versus human intelligence; it is American satellites (and other technical methods of collecting information) compared to the rest of the world. The incredible power of the information the United States collects via satellite inevitably overshadows human intelligence no matter what happens.
Technical collection does have limitations, but the intelligence community knows exactly what those limitations are. There is little ambiguity involved with technical sources. Yes, the intelligence community should devote more attention to human intelligence—and it should also devote more attention to improving its spy satellites as well. But we should also all be thinner and all babies should be raised in a happy home. Wishing for a better world will not make it happen unless we have a clear plan on how to get from our current state to that utopian ideal.
The real neglected humans
There is one criticism of humans and satellite intelligence that holds greater validity, but which is rarely voiced in the popular press. This is the charge that the intelligence community has not devoted enough assets to interpreting the photography. Last year that charge was leveled at the National Reconnaissance Office’s planned Future Imagery Architecture (FIA), which is supposed to consist of a constellation of satellites providing data from around the globe nearly continuously. The intelligence community was apparently more focused on the sexy aspect of the orbiting hardware than the less sexy subjects of the ground equipment and software needed to fully exploit it. The imagery analysis community also took a major hit in the mid-1990s with the merging of the National Photographic Interpretation Center (NPIC) and the Defense Mapping Agency (DMA) into the National Imagery and Mapping Agency (NIMA), which recently changed its name to the National Geospatial-Intelligence Agency (NGA). Amidst all the efforts to rearrange acronyms and bureaucratic flow charts, few people outside this insular community realized that many CIA imagery analysts took early retirement and left, resulting in fewer people to look at the photographs.
It is also worth noting that this claim that the analysts need more help is not new. A 1965 report on the National Photographic Interpretation Center by the CIA Inspector General concluded that the photo-interpreter population had been increasing at roughly three percent per year for the past five years—at the same time that the amount of imagery that had to be processed was doubling every year. The analysts needed more help, then and now.
If there is a real humint versus imint debate, it is over the amount of money spent on interpreting the deluge of information that the satellites provide. However, the claim that the American intelligence community relies too much upon orbiting iron and not enough on James Bond is a modern myth that deserves to be retired with extreme prejudice.