8:35 a.m.
Pilot: Alright, so the plan is, man, uh, we’re going to watch this thing go down and when they Winchester [run out of ammunition] we can play cleanup.
Sensor: Initial plan: without seeing how they break up, follow the largest group.
Pilot: Yeah, sounds good. When it all comes down, if everybody is running in their separate direction, I don’t care if you just follow one guy, you know like whatever you decide to do, I’m with you on it … as long as you keep somebody that we can shoot in the field of view I’m happy.
The crew was now making final preparations for the attack, arming the missile and going through the final checklist. The sensor operator reminded his intelligence colleague to focus on the business at hand.
8:45 a.m.
Sensor: Hey, MC.
Mission intelligence controller: Yes?
Sensor: Remember, Kill Chain!
MIC: Will do.
The first missile from the lead helicopter scored a direct hit on the pickup, instantly killing eleven passengers. The two following SUVs jerked to a halt, and the passengers began frantically to scramble out. The second missile hit the rearmost vehicle, but in the engine block, which absorbed enough of the blast to allow some of the passengers to escape. Four died immediately. The third missile missed the middle SUV, barely, with the blast blowing out the rear window as passengers bailed out. As a matter of routine, the attackers pursued these squirters, their word for people fleeing a strike, with 2.75” rockets, though all of these missed.
Then someone noticed something strange. The people who had escaped were not running.
8:52 a.m.
Sensor: That’s weird.
Pilot: Can’t tell what the fuck they’re doing.
Safety observer: Are they wearing burqas?
Sensor: That’s what it looks like.
Pilot: They were all PIDed as males. No females in the group.
Sensor: That guy looks like he’s wearing jewelry and stuff like a girl, but he ain’t … if he’s a girl, he’s a big one.
Despite the sensor operator’s hopeful theory, these were not Taliban in drag but women who had scrambled out and were waving their brightly colored scarves at the circling helicopters, which eventually ceased fire. Twenty-three people had been killed, including two boys, Daoud, three years old, and Murtaza, four. Eight men, one woman, and three children aged between five and fourteen were wounded, many of them severely.
9:10 a.m.
Mission intelligence coordinator: Screener said there weren’t any women earlier.
Sensor: What are those? They were in the middle vehicle.
Mission intelligence coordinator: Women and children.
The conversation in the Nevada trailer was losing its previously jaunty tone, as MAMs became mothers, and adolescents turned back into children.
9:15 a.m.
Pilot: It looks like, uh, one of those in the, uh, bright garb may be carrying a child as well.
Sensor: Younger than an adolescent to me.
Safety observer: Well …
Safety observer: No way to tell, man.
Sensor: No way to tell from here.
Soon afterward the Predator turned and flew away ahead of bad weather that was moving in from the west.
Even as the wreckage burned and shell-shocked survivors stumbled about, news was beginning to spread. Local villagers were soon on the scene, and within an hour Taliban radios were broadcasting word that “forty to fifty civilians” had been killed by an American air strike. By early afternoon, the reports had reached the Palace, the crenellated nineteenth-century fortress in the middle of Kabul that housed President Hamid Karzai. Meanwhile, U.S. military communications were proving rather less efficient.
The sudden, silent, flash of the first missile that incinerated the pickup and passengers on their screens caught most of the spectators in Afghanistan and the United States entirely by surprise. The intricate network of observation, control, and communication linking the myriad headquarters and intelligence centers stretching between Nevada and Kabul had somehow failed to alert participants—other than the crews actually pulling or preparing to pull the triggers—that events had reached their natural conclusion, and people were about to die. Then, even when it was almost immediately clear that things had not gone according to plan, the news moved at glacial speed through the U.S. command system. Messages rumbled back and forth between different headquarters regarding BOG (boots on the ground), meaning sending someone to have a close-up look at the scene for BDA (battle damage assessment).
Eventually helicopters were sent to bring the raiding party itself to the site where the dead bodies, or at least those that were intact, had been laid out by villagers who had flocked to the scene. The captain, according to a brother officer, was in a state of panic, searching fruitlessly for a weapon, anything, that would justify this as a legitimate target. “He wasn’t finding anything. I think it overwhelmed him.” Special Operations Task Force headquarters meanwhile told him “not to second-guess yourself; we’ll figure it out later.”
The captain was not the only officer to panic. Despite the services of a multibillion-dollar system of intelligence and communication, it took twelve hours for news that the U.S. had killed twenty-three civilians to make its way up the chain. Despite confirmation from the helicopter crews, the Predator team, and the troops that arrived on the scene, successive layers of Special Operations commanders refused to report CIVCAS (civilian casualties). Bizarrely, the technology was less efficient than the Taliban’s. With the inflated volume of traffic, emails were taking four and a half hours to move through the classified system from Kandahar to Kabul.
Only when surgeons at a Dutch military hospital talked to their U.S. counterparts about the wounded civilians that had just been admitted was the truth officially disclosed, but by that time, anyone in Afghanistan with a radio already knew. At the time, Stanley McChrystal, the U.S. and allied commander, was laboring to garner support among Afghans by restricting airstrikes in an effort to reduce civilian casualties. He was not pleased to hear the belated reports from Uruzgan, and raced over to President Karzai’s palace to tender his apologies. “I express my deepest, heartfelt condolences to the victims and their families. We all share in their grief,” he declared on Afghan television two days later. “I have made it clear to our forces that we are here to protect the Afghan people. I pledge to strengthen our efforts to regain your trust to build a brighter future for all Afghans.”
Families of the dead ultimately received $5,000 each, plus one goat.
McChrystal meanwhile appointed a senior officer who was also an old friend, army Major General Timothy McHale, to lead an investigation to determine exactly what had happened and why. McHale’s first act was to fly to the remote hospital where the wounded were being treated and meet the victims, among them a six-year-old boy, the same age as his own son, Riley, whose leg had just been amputated. “That really shook me up,” he told me later.
McHale, a logistics specialist appointed to command the entire supply effort for the U.S. expeditionary force, had only been in Afghanistan a matter of weeks., Now he quickly recruited a small but well-connected team of officers to help him explore the strange and, to him, unknown worlds of ISR (intelligence, surveillance, and reconnaissance) and DCGS (distributed common ground systems), not to mention Special Operations. Untrammeled by institutional connections to these organizations, he was unafraid to ask hard questions, an attitude that clearly upset some of the interviewees. The Predator squadron commander at Creech, for example, objected to McHale’s apparent impression that his crew “were out to employ weapons no matter what.” McHale tartly responded that according to the transcript of their chatter, the crew had stated exactly that intention “about fourteen times.” Furthermore, he pointed out, “You have a sensor operator whose response to a call out of children is ‘bullshit.’ Do you think he is likely to be focusing on potential for children or is he only looking for weapons or
trying to confirm that this is a target?”
Within six weeks McHale and his staff had interviewed over fifty witnesses in Afghanistan, Nevada, and Florida, creating in the process a hand-drawn time line of the events that ultimately stretched for sixty-six feet around the four walls of the hangar he had commandeered for his office. He delivered a withering report that described the Special Operations headquarters responsible for Operation Noble Justice as “ineffective,” while reserving his deepest scorn for the Predator crew, characterizing them as “almost juvenile in their desire to engage the targets,” and recommended that the air force conduct its own investigation of the crew’s “unprofessional conduct.”
The investigation’s interviews, transcribed and included in the report, track McHale’s exploration of the tangled links in the kill chain, much of which came as a revelation to the general, despite a lifetime in the service. Thus he was astonished to learn from the Florida team in their $750 million station that they had no means of communicating with the men on the ground in Afghanistan, relying on the mission intelligence coordinator in Nevada to pass on their information. “We cannot hear what he is saying,” one of the Florida staff told McHale, “so we hope that he is providing the best information possible.” The chief screener, an intelligence professional who supposedly had been trained to make lethal judgments on the basis of her observations, provided insight into her training in cultural awareness when she recalled how the vehicles had “stopped and a large group of MAMs began to get water, wash, and pray. To us that is very suspicious because we are taught that they do this before an attack.” Several hundred million non-Taliban Muslims also wash and pray every morning, but the little party’s ablutions had fed into the pattern already established by the flashing headlights and anonymous radio summons to the “Mujaheddin.”
The general eagerness in Nevada to “go kinetic” had done the rest. A safety officer, present to advise the Predator crew when the attack seemed imminent, summed up the prevailing attitude at Creech in a candid admission: “Well, to be honest sir, everyone around here, it’s like ‘Top Gun’: everyone has the desire to do our job, employ weapons against the enemy.”
McHale’s voyage of discovery as chronicled in the interviews, transcripts, and conclusions of his 2010 report not only retraced what he saw as a saga of bloody-minded incompetence and confusion but also revealed something more profound. The technological architecture in which the assorted participants operated was a tribute to the notion that if it is possible to see everything, it is possible to know everything and therefore automate the process of empirical deduction. Technology had supposedly made it possible to see down through the dark from almost three miles up and count the passengers inside a moving vehicle as well as any weapons they might be carrying. Technology enabled these images to move around the world for multiple viewers to assess and draw their own conclusions. Finally, it could be taken for granted that each target required only a single shot. In sequence, it was a very efficient kill chain.
But however miraculous the technology, the information it delivered was inevitably ambiguous (“Was that a fucking rifle?”) partly because, contrary to popular belief, the imagery delivered by ISR (the overworked acronym for intelligence, surveillance, reconnaissance) pictures tends to be fuzzy. Quizzing one Special Forces major as to why he had been slow to report the casualties despite the pictures coming from the drone, McHale remarked, “Your ISR knows there are civilians there…” To which the officer responded exasperatedly: “The ISR? Literally, look at this rug right here sir, that’s what an ISR looks like.”
Inevitably, everyone involved tried to clarify the ambiguity by shaping the information to fit a predetermined pattern, in this case that of hostile Taliban. Figures gathered on a riverbank at dawn to pray? “That’s what [the Taliban] do,” ergo they’re going to do something nefarious. For what other reason would three cars be driving away from the friendly force? Given the pervasive concentration throughout the U.S. war effort on the importance of HVI (high-value individuals), the ground force commander was quick to assume that this was an “exfill”—an assisted escape—of a high-ranking Taliban commander. Subsequently he accepted the alternative explanation of the “flanking maneuver,” equally well fit to the cast-iron assumption that the little convoy fitted the pattern of a threat. The complexity of this system, especially given its widely dispersed components (Nevada, Florida, Afghanistan), made it even harder for the people involved to adapt to changing reality. Instead, reality was adjusted to fit the predetermined pattern. This, in other words, was a “signature strike” in which the victims were targeted solely on the basis of their behavior.
Nor did it help that the system came with its own language—MAM (military-age male) for man, PID (positive identification) for see, TIC (troops in contact) for coming under fire—imposing its own framework. A military-age male, after all, is almost self-evidently a legitimate target, whereas a man might well be an innocent civilian. Officially fostered as a means of succinct, precise communication, the language adapted and divided, with different meanings for different people. So PID, for example, had a different definition depending on whether someone was in Florida, Nevada, or Afghanistan. Everyone had different notions of what adolescent meant and whether it was OK to kill one.
As recommended by McHale, the air force did indeed hold an investigation. Conducted by a major general, it concluded that the Predator crew had perhaps “clouded the picture on adolescents” but laid much of the blame on the special operations command for failing to supervise the operation. Neither of the reports highlighted the statement of a Special Forces sergeant from the ground force, a veteran of seven tours in Afghanistan, including three in Uruzgan, who lodged a protest against the system of complex technology embodied in the kill chain:
Looking at the video afterwards, someone was saying when the vehicles stopped, the (passengers) were praying. Someone said there might be people pulling security. When I looked at the video they could also have been taking a piss. Whoever was viewing the video real-time, maybe they needed a little more tactical experience. It needs to be someone that knows the culture of the people. If I can say anything, they just need to be familiar with what they are looking at.
But the system had not been built to work that way, not in a long time.
2
WIRING THE JUNGLE
Twenty years after the last bombs had fallen, the So Tri, an indigenous group who had lived in the remote wilderness of southeastern Laos for centuries, still didn’t know who had bombed them. For nine years, high explosives of all shapes and sizes had rained down out of the sky, killing men, women, and children and obliterating their homes and much of the old forest. The survivors had retreated deep into the mountains, hiding in underground shelters to stay alive. When the bombing finally stopped they came back and rebuilt their villages along the muddy trail they called the war road. Cluster-bomb casings dug from under the bushy bamboo that had replaced the forest were ideal as stilts to support their houses. The yellow bomblets could be turned into oil lamps, though some of them would still occasionally explode. Asked by a visitor in 1994 who it was that had bombed them over and over for all those years, the Tri laughed and shrugged: “The enemy.” Asked who the enemy was, they laughed louder and replied, “We don’t know.” When told that the bombs had come from the United States, they expressed thanks for the information, grateful to have the mystery solved at last.
It would have been harder to explain that at its heart the enemy had been a machine. A massive computer hundreds of miles away, prompted by devices hidden in the forest that were designed to detect the sound, movement, and even smell of humans and their vehicles, had directed when and where the bombs should land. Unwittingly, the So Tri had hosted the world’s first automated battlefield. Grounded on an unswerving faith that the vagaries of conflict can be overcome by technology, this half-forgotten project was the precursor of the drone wars that America would fight in the twenty-first century.
The scheme had been conceived far away on the east coast of the United States, in a leafy suburb of Boston, Massachusetts. Here, in 1966, Dana Hall, a prep school for girls, had been selected as the venue for the annual summer get-together of an elite and very secret group. Known as the Jasons, they were eminent scientists and scholars, most of whom were graduates of the Manhattan Project that designed and built the first atomic bomb. These men were accustomed to deploying their intellects to assist the U.S.government with the most fundamental and secret issues of national security, especially with regard to nuclear warfare. George Kistiakowski, an acerbic Russian-born physicist, had helped develop the atom bomb and gone on to be President Eisenhower’s science adviser. Carl Kaysen, who had helped plan bombing targets in World War II, held high rank in the Kennedy White House and at one point had urged a preemptive nuclear attack on the Soviets, providing precise calculations on likely casualties. The hyperactive Richard Garwin had worked on the design of the very first hydrogen bomb, tested in 1952, and now enjoyed a lofty position as senior scientist with the IBM Corporation. Many among the group were or would become Nobel Prize winners, and their ideas were assured of respectful attention at the highest levels.
Given their background in nuclear weapons, it was natural that the initial topics to which the Jasons were asked to address their intellects would involve nuclear warfare, specifically the problems of defense against intercontinental ballistic missiles as they came over the North Pole from the Soviet Union. It was assumed that with the right radars, enough computing power, and suitable interceptors, it would be possible to track and shoot down the missiles before they reached the United States. Possible solutions—the Jasons at one point suggested interceptor lasers—could never be realistically tested, so the problem remained pleasingly abstract, a rich field for abstruse technological speculation and, not least, a lucrative source of contracts for corporations such as IBM.
Kill Chain: The Rise of the High-Tech Assassins Page 2