The New Spymasters (36 page)

But how was it that the man who was killed was living openly in Kabul and Afghan government officials said he was innocent? What had convinced him? ‘Very precise intelligence that tells us exactly what he was doing when he was in Kabul, and exactly what he was doing up there. So again, there is not a question about this one, with respect.'
17

The truth, as all outsiders who investigated found, was different. It showed that US intelligence was not only questionable but – with respect – completely wrong. It revealed the Ant, Zabet Amanullah, as the Taliban double agent who wasn't.

When Rahman was arrested, he really did have an uncle who was a member of the Taliban and the deputy shadow governor of Takhar. (The Taliban appointed ‘shadow' administrators for each province in Afghanistan.) His name was Mohamed Aalem. He was forty-nine and the son of a well-known commander who had been killed in the jihad against the Soviets. Aalem, like most rebels, had adopted a nom de guerre in the war against NATO. His alias was ‘Mohamed Amin'. In short, he was the man the Americans had been looking for. All of his biography, including the names of his family and the fact that he had a home near Peshawar in Pakistan, matched the biography that US officials provided of their target.
18

Amanullah and Aalem (aka Mohamed Amin) were chalk and cheese. They were different people. Both were ethnic Uzbeks from Takhar Province, but while Amanullah came from Darqad District, Aalem came from sixty miles away in Kalfagan. The former had been living with his wife in Kabul, while the other still lived with his wife near Peshawar. One local elder in Kalfagan, Haji Khair Mohamed, confirmed that he knew Aalem and his nephew well. The nephew was in prison, he said, adding unprompted that Aalem was the Taliban's ‘deputy governor for Takhar province'.
19
None of this fitted any profile of Amanullah. And, despite what the US thought, Amanullah was not an alias. He was a famous man, a local hero. But these details were far too trivial for the mighty beast of US intelligence to know about.

*   *   *

Looking back, those who inquired into Amanullah's death concluded that he and Amin had been conflated accidentally. Perhaps some contact between the real Mohamed Amin and Zabet Amanullah was confused by eavesdroppers, so that Amin was imagined to be making the call, not receiving it. There was no doubt the Americans had recorded conversations involving someone who was a Taliban commander plotting an attack, but without access to their secret records, no one could be sure who exactly they were listening to on which phone at the time.

A few months after Amanullah was killed, I travelled round northern Afghanistan for a couple of days with the regional police chief, General Mohamed Daud. He was not a friend of the Taliban, but at the same time he knew them personally. Our days together were punctuated by taunting mobile phone calls back and forth between Daud and his enemy. Without context, without an understanding of the nuances of this man's relationship with the Taliban, someone who was tracing the Taliban commander's contacts could have mistakenly taken Daud to be a friend. In fact, he was killed by the Taliban soon after I last saw him, by a suicide bomber in Takhar, of all places.
20

Without access to secret records of the Amanullah case, no one can be entirely sure where the errors crept in. Michael Semple was convinced, though, that ‘one way or another some kind of blunder was made here. It's a classic case of somebody, who has legitimate reason for being in contact with someone designated as a terrorist, getting treated as “the terrorist”.'
21

*   *   *

In March 2011, Semple left his home on a farm outside Islamabad and made the familiar drive to the frontier city of Peshawar. With him were two trusted friends who witnessed what happened next. He was greeted in a hotel room by a middle-aged man with a black turban. The man showed him his Pakistani refugee ID card. It read ‘Mohamed Amin'. This was his alias, but he also had other papers that confirmed his true identity – Mohamed Aalem. Talking to Semple, he said he was aware of the NATO strike on 2 September. He was not surprised to be targeted. ‘This is war!' he said. But, just like Mark Twain, he noted that reports of his death had been greatly exaggerated.

Aalem said he had been promoted since the attack and was no longer the deputy governor. He also said that the Taliban had studied the incident and concluded that perhaps it had been a mistake by NATO signals intelligence. More importantly, Aalem added, a local informant with a grudge had simply told NATO that he and Amanullah were the same person and had then given them Amanullah's phone number. ‘This is not an isolated case,' he said.

Semple liked Aalem. He struck him as a ‘classic example' of someone the US had been targeting with the kill/capture campaign. ‘It's a tragedy that all we can do is kill such people, because he is a good Afghan.'

*   *   *

In May 2011, back in England, my phone rang. It was a call from Afghanistan, from someone who had been involved in the operation to kill Amanullah. I had to give JSOC full credit. On three occasions, I had been granted top-level access to ask questions about this with US officials. People involved in JSOC were smart, highly aware of the complexities of the Afghan environment, where the enemy would be eating, drinking tea and sleeping with the same population that the US was there to protect.

The official now calling appeared candid about some of the US intelligence gaps. He admitted, ‘We never disputed it was an election convoy. We just didn't know it [at the time],' he said. When asked about how the air strike was carried out, he also agreed that innocents might have died. ‘I accept a possibility that some of the people killed in the convoy were not combatants. Vehicles were each filled with armed men.'

But even now, after hearing what Semple had discovered, there was no trace of doubt that they had killed the right person. The incident had been reviewed again and again. An analyst who had checked and rechecked reported back, ‘There is absolutely no doubt in my mind that Mohamed Amin is one and the same person as the person who is known as Zabet Amanullah.' What about the man in Pakistan who was alive and said he was Mohamed Amin, then? The official replied, ‘Whoever this is in Pakistan, he is welcome to come and talk to us! We are dealing with an enemy that can create personae with ease. They can travel at will back and forth across the border with false papers.'

The American on the phone was perfectly reasonable and aware of all the contradictions, but clearly had total faith in the intelligence machine: ‘I realize that I am predisposed to believe the intelligence we had. But we really have looked to see if it's possible we blew this. There is no indication at all we can find we blew this.'

*   *   *

As in so many intelligence stories, it was impossible to be definitive here. JSOC never revealed all its sources. It is possible they also had some kind of human intelligence, an agent who gave them information that convinced them they had the right person. US sources certainly insisted that they had some ‘human intelligence'. But they were clearly missing a good spy, someone inside the network they were targeting who could have cut through all the confusion and told them who Mohamed Amin was, what his role was and where to find him. It cannot be proved that Zabet Amanullah was living a completely blameless life. Whatever the recollection of his friends, proving a negative like innocence is logically impossible. What can be proved, however, is that Zabet Amanullah and Mohamed Amin were two different people. Despite the assurances of highly placed figures in the US military, including the commander of all US and NATO forces in the war, and regardless of the outcome of internal scrutiny, the intelligence machine had been shown to be flawed.

Amanullah's assassination did not demonstrate that the methods used by the military in modern warfare – as exemplified by JSOC's campaign – were wrong. They regularly found and captured or killed the targets they intended. The principle of focusing everything on to a single target and the methods of network analysis based on tracking and analysing phone records, combined with prisoner interrogations, were generally solid. In the months ahead, they would lead US intelligence to the highest of targets. But these modern technical-based methods did not always work; it was easy to be confused by the data and make too many wrong assumptions.

As the campaign matured, officers in Special Forces who led the hard edge of the war in Afghanistan, as well as soldiers in regular battalions on the ground, all became wiser to their game and the nuances of the local environment. They were increasingly aware of tribal rivalries and how biased local informants could be. But often that made HUMINT look even less appealing.

As the real world ‘outside' began to seem ever more complex, it was ever more tempting to fall back on the certainties that non-human sources seemed to provide. When the CIA or SIS gave tips to the military, such as which village a Taliban group was hiding in, those involved recount how ‘before dropping the bomb' the military very wisely verified this human intelligence by technical means, such as seeing if the Taliban's mobile phones were in that village. But the reverse was not always true: there was not always solid human reporting to back a technical find. The new machines of warfare – the combination of technology and all-encompassing surveillance – were so intoxicating they appeared to blind many of their users, including well-meaning and highly intelligent people, to their limitations.

The problem was not the failure of particular technologies or methods. The glitches that caused Amanullah's misidentification may have been fixed the day after he was killed. The issue was over-confidence in the idea of technology itself: the infectious belief that somehow science and computation could overcome the insuperable problem of operating in a baffling and dangerous foreign environment.

The absence of sufficient HUMINT in a military campaign that depended on excellent intelligence had manifested itself in the wider battle to win over and protect the population, a campaign in which NATO had constantly made alliances with the wrong people. The same intelligence gap was also apparent in the narrow tactical campaign when, despite great efforts to avoid it, too often the wrong people were killed. In both cases the military were simply failing to work out who the real enemy was.

If there was too much faith in machines, what was the alternative in modern warfare? Was the need for more spies providing concrete secret intelligence – a ‘man on the rock' in each Taliban lair – or simply for more engagement that, addressing a wider intelligence failure, provided deeper context?

Both, in fact, were lacking. But of the two, the bigger gap was in strategic understanding, of the enemy and the wider population. Using spies to get more Taliban secrets to kill or capture more of the Taliban leadership was not going to solve the problem. Although fuelled by outside intervention, the Afghan War was a rebellion, which is to say a political and military conflict. However much the military might batter the rebels, they should have been asking whether the causes of the rebellion were being addressed and, valley by valley, whether the intervention of foreign troops was genuinely acting as a force for good. Answering these questions required an acute political awareness that went far beyond ordinary espionage.

Nevertheless, in practice there was a blurred line between the work of secret intelligence sources and the gathering of ordinary contacts and common-sense information. It is a truism (and amusing to observe in the field) that professional spymasters tend to dress up rather ordinary contacts as their ‘agents'. And since the enemy, the Taliban, lived among the people and were of the people, plenty of ordinary people knew specific secret information, such as the make-up of a Taliban group and where they were hiding. Moreover, the reverse was also true. Since very few Afghans were really prepared to be ‘recruited' as loyal agents for a foreign intelligence service, the best of agents might frustratingly provide little concrete intelligence, but only be useful in providing a broad overview of the conflict. In the circumstances, some argued that the best way of collecting intelligence on the enemy was to be open and do something as simple as picking up the phone to them, or joining them somewhere neutral for a cup of tea.

Such efforts were more like discreet diplomacy than espionage. But in Afghanistan, and in any dangerous or difficult place where contacts with hostile groups were politically sensitive, such engagements were becoming an essential part of the work of a modern spymaster.

Chapter 10

The Peacemaker Spy

‘The successful Field Officers will be generally found to have three important characteristics. They will be personalities in their own right. They will have humanity and a capacity for friendship and they will have a sense of humour'