Kill Chain: The Rise of the High-Tech Assassins Read online

  More hawkish factions were naturally gratified by the vindication of their strategic projections in combat. In September 1999, George W. Bush, then a presidential candidate, introduced his defense program at the Citadel military school in Charleston, South Carolina. Richard Armitage, the former Andrew Marshall subordinate who had cowritten the report on transforming national defense, wrote the speech. “Our forces in the next century must be agile, lethal, readily deployable,” said the future president, pledging to “begin creating the military of the next century.… Our military must be able to identify targets by a variety of means, then be able to destroy those targets almost instantly. We must be able to strike from across the world with pinpoint accuracy … with unmanned systems.” The opportunity to do this, he stressed, had been created by “a revolution in the technology of war. Power is increasingly defined, not by mass or size, but by mobility and swiftness. Influence is measured in information, safety is gained in stealth, and force is projected on the long arc of precision-guided weapons.” The lucrative technology dreams of William Westmoreland, John Foster, William Perry, and those they spoke for had now been endorsed at the highest level.

  One underpublicized feature of the Kosovo war had especially ominous portents for the future. There had been less of the public bloodlust to “go after” the enemy leader than there had been during the Iraq War. Officially, the highest-priority targets were the enemy’s “command-and-control” facilities. But, as one former targeteer remarked to me, “that could mean any place with a phone.” So Slobodan Milošević’s personal residence was duly destroyed. “We would have been happy to get him,” I was told by one former intelligence analyst who had been assigned to a “high-value target cell,” a new phenomenon in U.S. intelligence agencies dedicated to tracking the location of high-ranking human targets on an hour-by-hour basis. Assassination, officially forbidden and always denied, was still in the shadows but edging ever closer toward public respectability.

  Sharing in the glory of Operation Allied Force was the Predator, which Pentagon briefers extolled as a “CNN in the sky” that “enables us to see things in the battle space in a more human way … to use the unmanned vehicle for forward air control, much more efficiently and at much lower risk” than would be the case with manned aircraft. Just as carefully selected video clips transmitted from optically guided bombs and missiles had thrilled audiences during the 1991 Gulf War, snippets of Predator video footage “looking like it was shot from the roof of a fifteen-story building” now performed the same function. As one officer told a reporter, “Kosovo showed that UAVs are perhaps even more useful and can have more missions and roles than we may have thought.” Jumper himself excitedly reported to Congress how “toward the end of the war, we equipped the Predator with a laser so that it could place a beam on a target—this identified it so a loitering strike aircraft could destroy it … we developed a capability with great potential for rapid targeting.”

  Such enthusiastic hype, earnestly expressed, is traditional in high-technology defense programs, as demonstrated by the triumphant PR successes of the stealth aircraft and guided-missile systems assiduously promoted to the public during and after the 1991 Gulf War. As we have seen, the actual performance of these technologies was not quite as advertised: stealth planes were not invisible to radar, and precision missiles did not unerringly destroy their targets. It should therefore come as little surprise that the true story of Predator performance in the Kosovo War followed the same path.

  Apart from that one incident hailed by Jumper in which an experimental drone laser had assisted in the destruction of a target—an empty barn—the Predators in Kosovo were concerned purely with reconnaissance. As yet unarmed, they beamed streaming video to the JSTARS radar planes, designated to sift information from their own radar scans, from intercepted communications, and from other intelligence sources, and then disseminate the results across the NATO command. In a significant step along the road to remotely controlling the battle via headquarters on three continents, much of this intelligence was transmitted in real time to U.S.-based staffs for analysis and then relayed back to Europe.

  Meanwhile, thanks to the same expansion in communications bandwidth that made drones themselves feasible, the generals and admirals running the war were spending much of their days conferring with each other and Washington via video link. When not talking to each other they could watch the drone videos as they were being streamed directly into their offices, inevitably encouraging them in the belief that they had a close-up understanding of the ongoing war. General Clark himself, according to officers on his staff, was fascinated with drone TV and amid his busy days spent many hours glued to the monitor in his office. The general and his micromanaging habits were not universally popular with his brother officers, who were happy to circulate the story of how, one day, he called General Michael Short, the U.S. three-star commanding the allied air fleets. “Hey, Mike,” said the supreme commander, “I’m sitting here at my desk watching the UAV feed on the monitor. When are you going to do something about those two Serb tanks sitting at the end of that bridge?”

  In fact, Clark may not even have been seeing any tanks at all. Despite all the high-level enthusiasm and the release of carefully chosen videos, the all-seeing eye in the sky didn’t really work very well. We know this because, while Washington was still echoing with those rapturous reports, an organization immune to technohype was taking a cold, hard look at Predator. They were not impressed by what they found.

  In the right hands, the director of the Office of Operational Test and Evaluation is the most unpopular person in the Pentagon bureaucracy. Traditionally, the services have cast a benign eye on the actual performance of weapons programs they have fostered and do not welcome independent assessments of whether or not they actually work. That the office exists at all owes a lot in inspiration to John Boyd, the air force colonel who had arrived at Task Force Alpha in the waning days of the Vietnam War, shot the wild dogs, and closed down the essentially futile multibillion-dollar operation. Returning from Southeast Asia, Boyd had begun extrapolating the lessons he had deduced from earlier experiences as a supremely successful fighter pilot into a general theory of conflict that would ultimately earn him the title of the American Sun Tzu, after the legendary Chinese strategist. At the core of his conclusions was the concept of the OODA (the acronym for observation, orientation, decision, and action) Loop, the repeating cycle through which each side in a conflict passes. In air combat, for example, pilots see an enemy, orient themselves (meaning they subconsciously process their observation based on prior combat experience, intelligence, training, etc.), decide what to do, act on that decision, observe the results of that action, and continue retracing the loop. History shows that those who could adapt to changing circumstances—the antagonists’ own maneuvers and countermaneuvers—by continually moving through this cycle faster than their adversaries would prevail. Thus Boyd discerned that the F-86 fighter he flew in the Korean War outfought Soviet MiGs because its bubble canopy allowed the pilot a more complete view, while the plane could also transition from one maneuver to another faster than a MiG (partly because its power-assisted controls were easier to shift quickly than the muscle-powered controls on the MiG).

  In applying his ideas to organizations, as opposed to one-man machines, Boyd found that the same principles applied and that the overarching need for rapid adaptability to changing circumstances had to be based on a system of command and control that was as simple and harmonious as possible. It was extremely dangerous for the higher commander to try to get involved in the rapid pace and details of the firefight and thereby lose his focus on and grasp of the overall battle. Above all, Boyd stressed the importance of the human, as opposed to the technological, factor in warfare. One of his favorite quotations was Napoleon’s: “In war, the moral is to the material as three is to one.”

  As we have seen in the examples of Task Force Alpha, the Gulf War, and Kosovo, the U.S. military believes ver
y strongly indeed in material, the more complex and technically ambitious the better. Thus Boyd’s ideas as well as his emphasis on personal integrity were most certainly not in harmony with the prevailing ideology. Nevertheless, since he applied those ideas with great skill, not to say rigor (once causing a general literally to faint with rage in the course of a telephone discussion), in bureaucratic combat inside the Pentagon, he achieved considerable success in chosen objectives.

  Though highly unpopular in the commanding heights of the defense establishment, Boyd’s ideas had attracted a growing following in the military, especially among junior officers, as well as in the press and in Congress, giving rise in the late 1970s to what became known as the “military reform movement.” This alliance mounted serial campaigns against costly weapons programs of dubious utility, and the customs and practices of the weapons-buying culture that produced and nurtured them, exposing which involved revelations from whistle-blowers prepared, in many cases, to risk or sacrifice their careers for the greater good. For the most part these efforts were eventually defeated by entrenched interests in the military-industrial complex, but the movement, which for a time enjoyed potent support in Congress, did succeed in creating the post of Director, Office of Operational Test and Evaluation, mandated by law to test new weapons systems as a corrective to the services’ sorry record of buying systems that worked badly or not at all when deployed. As noted, the office was not popular with the military or with defense contractors, mainly because it regularly disproved claims by contractors and their service sponsors regarding the efficacy of lavishly funded systems.

  Even as the smoke of the Balkan battlefields cleared and the Pentagon echoed with claims regarding the success of Predator, a gravel-voiced Floridian mathematician named Tom Christie was taking over as director of the testing office. Christie was a friend and longtime associate of Boyd’s, having worked closely with him in the 1960s formulating a theory of air-combat tactics that many years later became official air force doctrine. Analyzing weapons effects at the Air Force Armaments Center at Eglin Air Force Base in Florida, Christie had a front row seat as colleagues worked to implement the high command’s obsessive determination to destroy the Thanh Hoa Bridge in North Vietnam. “They even wanted to turn a B-47 (a strategic nuclear bomber) into a drone,” he told me. “They’d load it up with high explosives and fly it into the bridge. That never got anywhere, just like all the other crazy schemes they had.”

  Technically able and skilled in bureaucratic maneuver, Christie managed to advance through the ranks of air force and defense department officialdom. By 1995 he had become director of the Operational Evaluation Division at the Institute for Defense Analyses, the Pentagon’s semi-independent think tank. In this capacity he reviewed an air force report on the sterling qualities of the performance of the JSTARS surveillance plane in the Balkans. As an example, the report cited an operation in which the system had detected the movement of a Serb armored unit while it attempted to hide in a cemetery. “We had built this beautiful topographical map of that whole area,” Christie told me, “and we knew exactly where the JSTARS had been at any particular time. So we were able to show that at the time they said they had spotted the Serbs in the cemetery, they were on the other side of a fairly substantial mountain. Even the air force couldn’t claim the thing could see through a mountain.”

  When Christie took over the testing office, the very first system that came up for review was the Predator. So, in the mountains and desert that make up the vast Nellis Air Force Base complex in Nevada, Christie’s team put the machine through its paces.

  The tests, carried out over nine days, mostly in a corner of the Indian Springs drone airfield, which had been activated five years earlier to preempt the army, did not go well. In fact, they were a disaster. One of the weaknesses revealed by earlier tryouts in the Balkans had been the aircraft’s vulnerability to ice on its wings, a fatal condition. In response, the technicians at Big Safari had developed a “wet wing” that could theoretically de-ice the wings in flight, and two of the four machines consigned to the tests were so equipped. But they didn’t work, a failure, as the testers later reported, that prevented “transition through clouds.” In fact the plane could not land or take off in any kind of bad weather, “including any visible moisture such as rain, snow, ice, frost, or fog.”

  Assuming it did manage to take off and reach enemy territory, the plane had a variety of cameras for viewing the ground and detecting targets. One of these, a “day TV continuous zoom,” looked at a 300-yard-wide area of territory in daylight from 15,000 feet. A second, the “day TV spotter,” could see in greater detail, but only over a narrower area of 50 yards. An infrared (IR) camera enabled night vision, while a synthetic aperture radar made it possible to see through clouds. To grade the cameras, the testing team relied on the National Imagery Interpretation Rating Scale, which runs from 1, the ability to pick out a large aircraft, such as a Boeing 737, to 9, the ability to recognize a human face. Though the Predator cameras were supposed to score 6, “recognize supply dumps, identify vehicles” at a range of 6 miles, the day TV scored no better than 2.7.

  Overall, Predator could find less than a third of its targets. As the testers put it, “[W]hen all targets are considered, only 29 percent of the targets tasked for the 7 days [of testing] were imaged.” The “day TV” camera that scanned the landscape could never deliver a picture sharp enough to enable a viewer to tell the difference between a tank and a truck. The infrared camera could manage that feat a fifth of the time. The close-up day TV spotter could tell the difference between a friendly and enemy tank just over two-thirds of the time, but since it captured only a small patch of territory—again, like looking through a soda straw—it was impossible to tell where the tanks were. The infrared camera could detect, as the testers reported, “something versus nothing,” but could discern what that “something” was, if a tank or a truck, only 21 percent of the time. Much of the time the aircraft failed to reach the target area, because of bad weather, engine problems (exactly half of the test flights flown under combat conditions never completed the mission), or some other breakdown. One or another component vital to continuing the missions failed on average every 19.5 hours, while some other major system failed every 3 hours.

  It made no difference. The drone could relay pictures, in color, of the enemy landscape. In Vietnam, troops fighting on the ground had been supervised by ascending tiers of senior officers hovering high above the battlefield in their personal helicopters, trying to “fight vicariously through that frightened twenty-five-year-old down there beneath the tree canopy,” as one veteran later wrote. There were even stories of entire units hiding from micromanaging heli-borne commanders in the sky above. But now, when the weather was fine, four-star generals could take the role of junior squad leaders without having to leave the office and even help destroy targets themselves.

  “The air force had had this idea that the [Predator] could be used for detecting patterns on the battlefield,” an air force officer intimately involved with the program told me later. “It would send back the FMV [full motion video] from a wide area, and that would be compared with previous video so they could find changes—new units moving in or whatever. But the quality of the video was lousy, as those tests out at Nellis showed.”

  The testing team had been kept in ignorance of one aspect of the Predator program that was to prove all-important. While they monitored its progress, or lack of it, in the bleak Nevada fall weather, another team had been secretly working to “weaponize” the machine: to arm it with a missile so that it could kill people. The scheme, initially ordered by Jumper in June, had gained urgency following a series of Predator flights flown from a base in Uzbekistan on behalf of the CIA over eastern and southern Afghanistan in September and early October 2000. They had one objective: to find Osama bin Laden, who was a high-value target. An address receiving special scrutiny was a cluster of compounds east of Kandahar known as Tarnak Farms, thought to be
frequented by the infamous terrorist leader. Sure enough, on one particular day a Predator beamed back tantalizing images of a figure in white, surrounded by others in dark clothes, moving down a street in the compound.

  Those few seconds of footage had momentous consequences. Back at CIA headquarters in Langley, Virginia, officials eagerly interpreted the pictures as showing the six-foot-five-inch Bin Laden himself—the white-clad figure seemed taller than the others—walking to the mosque, surrounded by deferential bodyguards. Just as the muddy pictures from Kosovo had kept General Clark glued to his monitor, so the Tarnak video had a potent effect across Washington. CIA Director George Tenet, who had shown little interest in the drone program heretofore, was suddenly a convert, screening the video for President Clinton and his National Security Adviser Sandy Berger at the White House, declaiming enthusiastically about the drone’s capabilities to the intelligence committees on Capitol Hill.

  Yet, the closer one looks at those pictures, the less they reveal. The “tall man in white” is actually just a moving white dot surrounded by black dots. It takes imagination to read the black dots as “deferential” toward the white dot and more imagination to define them as bodyguards, since no weapons are visible. Blown up, the images do not reveal more information, but less. The white dot does not turn into a tall man with a beard but merely a fuzzy blur that becomes more indistinct the more it is magnified. Just as the Predator crew in the first chapter of this book interpreted a carload of praying Afghans as Taliban preparing an attack, so eager Washington policy makers saw what they wanted to see and proceeded accordingly. Had the drone only carried a missile, they surmised, the mastermind of the lethal attacks on the U.S. embassies in Africa and the USS Cole off the coast of Yemen could have been eliminated in a single stroke. The pressure to arm the Predator was overwhelming.