The New Spymasters

by Stephen Grey

  That evening, US Secretary of Defense Robert Gates happened to be in Afghanistan and he held a press conference with the country’s president, Hamid Karzai. When that day’s attack was mentioned, the president spoke bitterly. ‘Pro-democracy people should be distinguished from those who fight against democracy,’ he said. Gates responded, ‘This is the first I have heard that civilians have been killed and we will look into that.’7 But, ten days later, NATO issued a new statement reiterating that the right target had been struck, even though civilian casualties ‘could not be ruled out’.8 NATO also confirmed media reports that the target’s name was ‘Muhammad Amin’. This caused new confusion. Was this some kind of code name for Zabet Amanullah? Either Amin was another man among those in the convoy who died in the air strike or, if the US was to be believed, the Ant was a Taliban leader with two identities – in effect a double agent.

  As was often the case, NATO’s comments on the Takhar air strike emphasized the difference between how Afghans viewed their country and how it looked to foreign eyes. Even Afghan officials who dealt with NATO and welcomed its presence in the country often concluded that, for all its high-tech wizardry, America’s spy machinery was rotten. Every day there were more raids and more strikes against the enemy. Sometimes, for propaganda reasons, it had suited President Karzai and others to criticize American air strikes, even when the Afghan government privately knew that the victims were probably Taliban fighters. But at other times, when it was obvious to them that the US was using bad intelligence to kill the wrong people, they were furious. The death of Amanullah epitomized those errors.

  * * *

  Intelligence about the enemy’s plans and disposition and about the zone of combat has always been essential to soldiers. But it took on an even greater importance in the war against the Taliban. This had begun in 2001, when the US invaded Afghanistan and toppled the Taliban’s regime. By the mid-2000s, the Taliban had regrouped. By the time of the Takhar strike nearly 100,000 American military personnel were deployed to the country, along with 40,000 other NATO-led foreign troops (including 9,000 Britons). This was more than the Soviet Army had there in the 1980s.9 By the end of 2010, over 2,200 troops in the US-led coalition had been killed.

  The conflict in Afghanistan was what the military call an unconventional, or asymmetric, war: Afghan government and coalition forces in uniform were fighting a Taliban that acted as irregular rebels, dressing as non-combatants, living secretly among the population, adopting guerrilla tactics of surprise ambushes and avoiding conventional battle. In military-speak, this was a classic insurgency. And although, historically, rebels tended to win such conflicts, the only known way of defeating them was by making use of super-precise intelligence. A successful counterinsurgency strategy was based as much on trying to separate and protect the population from the insurgents as it was on fighting them. For this to happen, intelligence was needed on who should be protected (friendly or neutral people) and who should be targeted (the enemy). This was hard because to foreign eyes they all looked alike and often lived together.

  Intelligence for NATO’s campaign came from the military’s own intelligence specialists – whether from intelligence officers working in frontline battalions or specialized military cadres, like the Defense Intelligence Agency (DIA) or the National Security Agency (NSA). They were assisted by deployments from the civilian secret services, primarily the CIA or SIS. These agencies handled particularly sensitive sources or specifically political agents, as well as dealing with Afghan spy agencies and conducting their own covert operations.

  As the violence intensified, both diplomats and secret service operatives based in Kabul faced increasing threats to their lives and, bound by strict health and safety rules, they were often restricted from going out and making their own contacts. What intelligence they did get from spies was mostly second-hand, the product of their liaising with and mentoring of local security forces. These included, as previously described, the various semi-private paramilitary groups the CIA ran directly, as well as Afghanistan’s own security service, the National Directorate of Security (NDS). ‘Even if we had people who had learned to speak like locals, we would never have looked like them,’ said one British intelligence officer. ‘There was a limit to what we could have done ourselves.’ The NDS had many faults (it sometimes tortured its prisoners, for instance), ‘but it had a network of sources nationwide. We could never have competed with that.’

  The trouble with having few spies of their own was that Western agencies were always vulnerable to being used to settle local feuds. It was common, for instance, for Afghans to relay tip-offs from someone in one clan that someone in another was linked to the Taliban or al-Qaeda. Many of Amanullah’s friends wondered if the Americans had been fed information by a particular local politician who, historically, had been a major rival of Amanullah’s family.

  Intelligence agencies were also aware of an inbuilt Afghan suspicion of foreigners or outsiders. Before British and other NATO troops began deploying in large numbers in southern Afghanistan in 2006, and got drawn into heavy fighting, SIS operatives had reconnoitred the area with Britain’s elite Special Air Service (SAS). The mission reported that at the time no insurgency existed but, given the population’s hatred of armed foreigners, there would be one if the army engaged.

  This specific warning was ignored, and so was the implication that basic intelligence about the instincts and allegiances of ordinary people in the countryside was at least as important as specific intelligence about who was a Taliban fighter or leader and where he was hiding (whether that was obtained using a spy or a radio intercept).

  One NATO intelligence chief, then Major General Michael Flynn, had revisited this weakness in early 2010 when he wrote that intelligence officers had focused so much effort on insurgent groups that ‘the vast intelligence apparatus is unable to answer fundamental questions about the environment in which US and allied forces operate and the people they seek to persuade’. US intelligence officers and analysts were ‘ignorant of local economics and landowners, hazy about who the powerbrokers are and how they might be influenced, incurious about the correlations between various development projects and the levels of cooperation among villagers, and disengaged from people in the best position to find answers’.10

  The gap Flynn had identified was human intelligence, but it was not the sort of high-level secret intelligence that could come only from a top spy. Rather, it was the sort of cultural understanding that ordinary dialogue with local people might have brought about.

  Military officers sometimes complained that their partners in the civilian intelligence agencies had become too ill-equipped and too bureaucratic to operate in a war zone. One former Western commander described offering an SIS officer a trip on a helicopter the next day to meet locals in a recently captured town. ‘Sorry, I’m not sure I can get the business case through London by then,’ the intelligence officer told him. This response seemed to sum up multiple problems.

  But as I witnessed while covering the war as a journalist and spending many days with frontline troops and commanders, over the years of Afghan engagement, intelligence gradually improved; the British and American armies devoted huge efforts to becoming more sensitive to the local human environment. But it was never enough and the improvements were from a very low base. For example, barely three dozen people in the entire British Army in the mid-2000s could speak fluent Pashto, the language of southern Afghanistan. They may have tried, but the military was not equipped to gather the intelligence it needed. And the realization of their deficiencies here came too late. While some intelligence officers would write off this kind of missing intelligence as ‘low-level atmospherics’ beyond their responsibilities, its absence was one of the reasons why the military campaign went wrong. The US and NATO had frequently blundered into one valley or another in cooperation with deeply unpopular warlords or corrupt government officials who were linked to a particular tribe. That had only antagonized other tribes
and strengthened the Taliban’s hand.

  While this wider picture about the terrain of battle was, at least at first, far too neglected, US intelligence agencies worked hard on helping their military develop its aggressive and innovative manhunt for top enemy commanders: people like, as the military believed, Amanullah. The object of what was called the ‘kill/capture campaign’ was to pummel the enemy by assassinating or capturing its leadership. The targets were to be identified and located by both spies and other human sources, as well as with data from video surveillance and the interception and tracking of phones and radios. As described above, all relevant information was combined in the ‘fusion centre’, which was designed to make different agencies work together effectively. First assembled at Balad airbase, north of Baghdad, the centres used makeshift buildings laid out as spokes around a central hub. At Balad, some began to call it the Death Star, and the name stuck as the same operation was moved to Afghanistan. The system had been pioneered in Iraq by a US Special Forces general, Stanley McChrystal. He had then been commander of the secretive Joint Special Operation Command (JSOC), which directed the activities of America’s elite Special Forces, the Navy Seals and Delta Force, supported by the Ranger Battalion and working also with Britain’s SAS and its Special Boat Service (SBS). McChrystal’s idea was to pull together almost every conceivable intelligence tool available to the US and focus on cracking a single objective. Its most successful implementation was in finding, locating and then killing the bloodthirsty al-Qaeda leader in Iraq, Abu Musab al-Zarqawi. Task Force 535 was just the latest cover name for JSOC’s forward headquarters and operations in Afghanistan.

  Everything about the Death Star approach was based on intelligence, but it also depended on speed. One key person involved said, ‘With targeted killings you either defeat or you shape the enemy … If you get the tempo high enough then it’s difficult for them to come up for air. It’s like hitting a boxer with body blows; it’s not a knockout but you stop them from breathing; you’re keeping them off-balance.’11

  This was war by raiding, and the tempo was maintained, not so much by kill operations like the one against Amanullah in Takhar, but by Special Forces ground raids at night to kill or capture prisoners. If the target surrendered, he would be taken back for interrogation and, even if he did not, his home would be searched for every type of material. Intelligence pulled together on one night could be used to launch another raid on the next. The soldiers looked not only for bigger items like laptops but also for ‘pocket litter’. This meant phones, SIM cards, notebooks, scraps of paper, anything that gave clues about the target’s network – who he was connected to. The McChrystal approach was all about tracing connections, using every available piece of information to move rapidly from one target to another. According to someone involved, ‘We have had decades of manhunting. We hunt individuals, but what’s changed is we have started to target entire networks.’

  The biggest source of HUMINT for the manhunt was prisoners. With the help of interpreters, JSOC got night-and-day access to question the enemy in jail cells close to their headquarters. The US asserted the right to run its own military prisons and only hand over prisoners to the Afghan authorities after their thorough debriefing. One visitor to the Death Star described hearing a live audio feed from an interrogation room being piped to his work station. ‘It was like listening in to the enemy’s mind. It was incredible,’ he said.

  The final major source of information – the key to what happened in Takhar – was technical intelligence: the constant interception and, as important, tracking of mobile phones and VHF radios, as well as visual surveillance of buildings, vehicles and gatherings of people by means of spy satellites, surveillance planes, helicopters and what had become a huge fleet of drones, each of which had several cameras.

  But there were flaws. One of the biggest was that, in order to persuade the special agencies to gather together and share all they knew, the headquarters had to be kept very secure and secret, and could only use the most elite of security-cleared soldiers and the minimum of outsiders.

  As he was discussing it, one US military officer involved in the Amanullah case tellingly made constant reference to the world ‘outside’. The war had divided people into insiders and outsiders who lived in parallel. Insiders like this officer lived within the ‘bubble’ of bases fortified by razor wire. When they did venture out it was usually to some other ‘secure location’ or, if not, in a posse of men armed to the teeth. Ordinary human interaction became impossible. They were cut off from real people.

  While this elite had access to tremendous technical tools with which to observe the world, all the secrecy and isolation stymied their ability to check and understand what they picked up. It was hard to look at a problem as a whole or understand the significance of certain elements. In intelligence-speak, nuggets of information tended to get lost in vast ‘silos’. Because everything was kept secret from the wider world, some basic false assumptions – obviously wrong to any man in the street – would never be challenged. And all this scientific espionage was also bewitching. Cool gadgets and smart techniques inspired awe and a confidence that was comparable to religious zeal. It defied good sense.

  And there was a further big problem: the absence of good spies. Some reliable secret agents among the Taliban could have made all the difference. But the tempo of JSOC operations made that difficult. Certainly, prisoners held at Bagram could be recruited, but the complex task of running such agents among insurgents in the field was a different matter and not to be tackled lightly. Yet in the absence of high-level secret information from human sources, it could be hard to challenge compelling, if misleading, intelligence from technical sources. Common sense dictated that Amanullah was innocent. But if his voice had been captured by secret interception and the words he spoke seemed suspicious, the Death Star would have needed a reliable source very close to his circle to exonerate him by explaining that what he said was innocent. Technical intelligence, unencumbered by coverage from human sources, could be dangerously persuasive. To find the truth, intrusive surveillance almost requires its mirror image: intrusive spying.

  * * *

  Obscure and remote as it was, the assassination of Zabet Amanullah in September 2010 caused a shock. He was widely known back in the Afghan capital, Kabul, including by some influential and well-connected people. It was their anger about his killing that motivated their efforts to discover how he was targeted. Their investigation has provided a unique window into the twenty-first-century intelligence machine.

  One person who knew Amanullah well was an Irishman named Michael Semple. He was one of those rare people who, by virtue of his work, straddled the worlds of the secret and of the ordinary, which in turn gave him some unique insights. He had come to the region twenty years earlier. Working for the UN and then the EU, he had gradually become involved in trying to foster political reconciliation in Afghanistan. As he made contact with every kind of political, military and religious group, he came to know the men of violence. In 2008, President Karzai had expelled him from the country for allegedly unauthorized contacts with the Taliban, but Semple carried on the same work from Pakistan, where most of the Taliban leadership lived. As an interlocutor, he also frequently crossed paths with Western military and secret intelligence.12

  What Semple noticed was just how often those in the secret services convinced themselves of false notions. And just such wrong-headed thinking had led to Amanullah’s killing. Semple had known the man for years and would not accept that he had been a secret Taliban commander. Even years before, when the Taliban ruled the country, Semple remembered that Amanullah had helped research the regime’s human rights abuses. And then, after the US invaded and the Taliban was toppled, the pair stayed in touch. Semple remembered introducing him to a delegation of British Members of Parliament in Peshawar one week in 2003.

  Since 2008, he said, Amanullah had lived peacefully in Kabul and ‘nobody would have considered him a Taliban looking
at him there’. If Amanullah had still been with the Taliban, he would not have been involved in the election campaign in Takhar, said Semple. It had meant ‘travelling village to village very publicly and giving speeches. Everybody saw that. So much of Zabet Amanullah’s life was in the public domain. And it had nothing whatsoever to do with insurgency.’

  Another of those who knew Amanullah was a former BBC foreign correspondent, Kate Clark, who had lived in Kabul under the Taliban and stayed on after. She left journalism to join an academic group called the Afghan Analysts Network, but she had not lost her detective instincts. She first met Amanullah two years before his death, in circumstances that convinced her he could not be an active fighter. He had described being tortured by Pakistan’s ISI for refusing to join the Taliban. Even if the US had been right and he had a secret role with the insurgency, why, she wondered, had they simply not arrested him at his Kabul home? She knew he had settled in Kabul, had bought a pharmacy and was studying English and computer science. After his death, she collected paperwork that proved it. The implication from US intelligence was that Amanullah was leading a dual life, that he was some sort of Taliban secret double agent. But then another explanation gradually dawned on her: US intelligence was not even aware of his home and life in the Afghan capital.

  Investigating doggedly, Clark used field research to establish that Amanullah was the only person of importance in the convoy, the target of the strike, and that he was definitely the person that NATO referred to as the IMU/Taliban leader ‘Mohamed Amin’. Using her contacts, she then pressed NATO commanders in Kabul for an explanation and finally established direct contact with some officers from JSOC (aka Task Force 535) who ran the operation.