by Stephen Grey
The knee-jerk response to cases like that of the ‘underpants bomber’, as he was called, was to collect and analyse yet more information. But that only guaranteed more information overload. As one technical expert warned, ‘The more data you collect, the more you struggle to process, interpret, and move it. The bad news is that an avalanche can bury you alive.’31 Intelligence collection was being overwhelmed by its own capabilities, but the same high-speed digitization capabilities that the NSA could so readily exploit were also the source of added real-world complexity, to some extent neutralizing the advantage gained. Digital financial transactions, for instance, meant money movements were easier to trace, but also that they were faster. The world itself was getting harder to read. As the director of NSA’s signals intelligence, Maureen Baginski, had explained in 2001, ‘You could literally stare for 25 years at the Soviet land mass and never have this kind of volume problem. They were slow, so it was okay if we were slow. Today, it’s volume, it’s velocity and it’s variety.’32
* * *
One of the biggest weaknesses of the digital manhunt was that those most susceptible to being tracked digitally were the innocent. They had no special reason to encrypt their emails or adopt false identities or anonymize their use of the Internet. As the CIA discovered, when they worked through papers and files seized at his compound, Osama bin Laden had stayed completely off the telephone and Internet grid. He had sent couriers dozens of miles away to transmit his emails from random public computers. My research into the CIA’s rendition programme showed that, at least in the early days after 2001, too often people had been wrongly labelled terrorist suspects because some overly simple link analysis had classed an innocent connection with a suspected militant as proof that that person was a militant too. An analysis of a terrorist’s phone calls might show many calls to another number; he might have been calling his girlfriend, who had no knowledge of his crime. But the key call to a terrorist associate might actually be made via another phone – a payphone, for instance. This was the ‘law of weak connection’: the weakest link might actually be the most important.
When I asked veteran intelligence officers about the quality of technical intelligence – particularly intercepts – over the years since the Second World War, most suggested that it went in waves. Over several decades, the CIA were sent a copy of every telegram in and out of the United States. All overseas phone lines were at one point tapped. There were years when interception had huge coverage. Then people found other ways to communicate, different codes and encryption, and legal restrictions were enacted by Congress. Some even argue that the expansion of the digital world has left the ultimate customer, the political leader and the security agency, with broadly the same level of secret intelligence, only collected now at vastly greater expense. ‘It’s just impossible to keep up,’ said one former CIA chief of clandestine operations, although he did not suggest not trying. But while that pessimism is justified when dealing with hard secret intelligence targets – those who try to conceal their secrets – the truth is that the modern citizen is easier to find and put under surveillance than ever before. What has also certainly changed, with advances in technology, is the ability of technical intelligence to work really well in hindsight – in reconstructing events and tracing known enemies. It remains, as always, much less good at looking forward and predicting new targets, new threats.
* * *
With all this tracking and technology, how have real spies fitted in?
In Britain, the intense study of travel plans and networks of suspected militants helped pinpoint targets for recruitment of agents. Of particular interest was anyone who had attended places where militants ran training camps (such as the Pakistani tribal areas or Somalia) or where there were active conflicts, such as in Syria. As Mike Sheehan, former counterterrorism coordinator at the NYPD, put it, ‘Connectivity back to the camps is the key to being operationally effective.’33
According to some youngsters who were approached, and their lawyers, one method used by MI5 to recruit was to discover some violation of immigration laws in a suspect’s family and then put pressure on the target to cooperate. MI5 has made use of new powers under the Terrorism Act 2000 to detain, question and search young British Muslims when they arrive back in the UK after a trip abroad. Although these individuals were not suspects, they said they were pressured to work as informants for MI5. Mohamed Nur, a 25-year-old community worker in Camden, north London, said he was visited by an MI5 officer and a policeman disguised as a postman. He told a newspaper, ‘The MI5 agent said, “Mohamed, if you do not work for us we will tell any foreign country you try to travel to that you are a suspected terrorist.”’34 His claim could not be verified, but within certain Muslim communities many youngsters certainly felt harassed. It was also arguable that, if useful intelligence could be garnered by such methods, then some ill-feeling was a price worth paying.
Security officials in Western agencies typically denied using blackmail, claiming it would backfire, would produce unreliable information and would be unethical. But as one veteran CIA case officer put it, ‘That’s all bullshit. We do what we have to do.’ He pointed out there was a distinction between a straight and unsubtle blackmail threat and a more nuanced exploitation of a weakness. A recruiter might, say, let the target realize he was aware of that target’s weakness (for example, that he had entered the country illegally), without uttering any explicit threat to expose him. Or, even better, once such weakness had been identified, the recruiter could try to offer himself as a solution to that problem (for example, offer to legalize his status). But, while less brutal, this was still a form of blackmail, he argued.
One former British intelligence officer assessed the pressure like this: ‘Faced with a huge indigestible mass of potential leads, coming mainly from technical sources, there is no time for the long-term patient cultivations of former times. So, yes, threats may be employed, and some of those doing the threatening will be incompetent, and, naturally, you will read of the approaches that failed and not of the ones that succeeded and, so far, keep us safe.’ One recruitment attempt by MI5 deserved particular scrutiny. In May 2013, after the brutal daylight murder of a British soldier, Lee Rigby, outside the Royal Artillery Barracks in Woolwich, south London, it emerged that at least one of the two men involved, Michael Adebolajo, was well known to British intelligence. A friend of his told the BBC that when Adebolajo returned from a trip to Kenya, he was ‘being harassed by MI5’. The friend, who was himself arrested after the interview, said, ‘His wording was, “They are bugging me; they won’t leave me alone” … He mentioned initially they wanted to ask him whether he knew certain individuals … But after him saying that he didn’t know these individuals and so forth, what he said is they asked him whether he would be interested in working for them.’35 Others confirmed he had complained previously about harassment.
Had pressure from British authorities contributed to Adebolajo’s murderous nature? There was insufficient information made public to allow anyone to judge. But Adebolajo was a committed jihadist and mentally unstable long before any MI5 recruitment attempt. Both attackers were found guilty of Rigby’s murder at trial in December 2013 and were later sentenced to life imprisonment.36
By getting agents in and among low-level operatives who came and went to training camps, there were some real successes in preventing attacks. With his operation in both Britain and Denmark, Morten Storm, the Danish ex-biker, showed the value of having agents in militant circles who could act as spotters, keeping an eye on those who either disappeared off for training or seemed to have a genuine desire to ‘turn operational’ and carry out an attack. Storm’s own intuition in spotting such would-be terrorists demonstrated the value of the human touch.
Now also working for the CIA, Storm was sent back to Yemen, where he had studied Islam, and he befriended the now-notorious Yemeni-American preacher Anwar Awlaki, who had become both an influential online preacher of al-Qaeda propaganda and a lea
der within the movement’s Yemen branch, al-Qaeda in the Arabian Peninsula (AQAP).
While running spies inside al-Qaeda is fraught with problems, Storm’s case showed how lateral thinking could overcome some of the issues, with a modern blend of the human touch and technical wizardry. One weakness the CIA exploited was that, away from his life as a militant, Awlaki was a normal man with physical and emotional needs. Storm was able to establish contact with Awlaki, but avoid getting so close that he (or the agencies running him) was compromised, by acting as a trustworthy but not entirely convinced supporter who was willing to send him or bring him out supplies. As recorded in videos and emails, Storm (and indirectly the CIA) served as a matchmaker, finding Awlaki a new wife who arrived bearing a suitcase that was – very helpfully – fitted out with a tracking device.
When that bag went astray, the CIA sent Storm back again, passing on more tracking devices in various disguises. None of those appear to have worked, but Storm believes Awlaki was finally killed by a drone strike after he led the CIA to a courier used by the cleric, handing the courier a USB stick with an inbuilt tracking device.
Storm finally went public with the story of his spy operations after a disagreement with the CIA about whether he should receive the financial reward offered for Awlaki’s death. He taped his final meeting with the agency with his iPhone. The CIA officer, ‘Michael’, explained that Storm’s mission had been one among many attempts to get Awlaki, and it was another that succeeded. According to Storm, ‘It’s like being on the field at the World Cup, you’re moving down the field and you’re in the position to score, the other guy could have passed it to you but he didn’t, he took the shot, he scores. And that’s that. That’s what happened.’37
But Storm also said that he came to understand why the CIA refused to acknowledge that he had led them to Awlaki: as he was a Danish asset, it would have meant that Danish intelligence had assisted with an assassination, something prohibited under Danish law. And this was a major problem with collaborations with the United States. Although Western spy agencies did have a common enemy, at the sharp end they took different approaches. SIS had previously made great efforts to persuade Storm to work for them while also trying to avoid any connection to targeted killings, which were prohibited under British law. But the CIA had offered Storm more money.
The other known successful recruitment of a spy inside AQAP was a British operation, run with the help of Saudi intelligence, that was able to thwart yet another attempt to blow up a passenger aircraft. According to intelligence officials, the plot was discovered by an agent run by Britain’s MI5. Recruited in the UK, the agent, who was of Saudi origin, had been given a UK passport and sent to language school in Yemen to follow in the footprints of Umar Farouk Abdulmutallab, the Nigerian ‘underpants bomber’. This new operative, a double agent, made it to the mountains of Shabwa, southern Yemen, where he penetrated the cell. He was finally sent on a mission with another underpants bomb, which he duly handed over to his handlers. The US then followed up with a series of drone strikes. This, for the agencies concerned, was nothing but good news. However, the British were furious when the existence of the agent emerged in a typical Washington farce. The Associated Press agency had learned of the agent from a source, later discovered to be an FBI contractor.38 They were prevailed upon to delay publication. But when, on 7 May 2012, they disclosed that a bomb plot had been disrupted John Brennan, Obama’s counterterrorism chief and a future CIA director, briefed several television pundits in advance that the plot had never really been a major threat because of ‘inside control’. This was correctly interpreted to refer to an agent. Richard Clarke, a former White House official, then told ABC Nightline, ‘The US government is saying it never came close because they had insider information, insider control, which implies that they had somebody on the inside who wasn’t going to let it happen.’39 All this led to an agent’s existence becoming public. Presumably, the agent was quickly retired and is living under protection.
In contrast to operations in Yemen and Somalia, the protection given by the ISI in Pakistan to militants and tighter security in frontier regions near the country’s border with Afghanistan meant there was less success in developing agent networks or running agents in the camps. The CIA tried to send as many operatives as possible into the country, and they also attached officers to the ISI base in Miranshah, North Waziristan, although they never left the base. This was not to say, however, that secret services got no agents into the camps. From time to time, they did, including some agents sent by the British through Peshawar.
Although US intelligence did get better at finding and striking in the tribal regions against senior leaders among the foreign militants – usually still labelled as al-Qaeda – they proved less successful at finding members of the Afghan Taliban who were hiding out. Nor could they find a US soldier, Private First Class Bowe Bergdahl, who was kidnapped in 2009 after wandering off his base and was held by the Haqqani faction of the Taliban, based in North Waziristan, Pakistan. (He was eventually released in a swap for Guantanamo Bay prisoners.)
According to one senior US intelligence official, the success of drone strikes had little to do with spies on the ground and much more to do with a combination of technical methods – mostly overhead surveillance and tracking phones – as well as secret operations led by JSOC to enlist the help of Taliban prisoners: ‘The intelligence is almost entirely coming from us; we get almost nothing from the Pakistanis. Yes, some direct HUMINT, but really it’s because we’ve been watching these places for years. When I say staring for years, I really mean staring for years now. We really know these places’ (his emphasis).
He said that interrogation techniques had been refined since the days of crude and harsh interrogations after 9/11: ‘We sit them down; it’s all completely different than it used to be.’ Apparently, 99 per cent of them would talk: ‘It’s like a survival experience. They’re so happy they’ve emerged from this whole experience alive and now they’re safe.’
Taliban who cooperated were given training and could be used to pinpoint precise compounds in Pakistan that militants used: ‘They can describe who is in every room in that Haqqani madrasa or whatever.’ Some even received training in surveillance technologies: ‘We teach them. They get lessons in how to understand overhead imagery. They start with stuff that’s irrelevant, just get taught the techniques. Then we show them things they should know about and they start saying, oh yeah, that’s the place we used. They talk us through it all. It works.’
This rather amazing operation to recruit members of the Taliban to help target drone strikes, which to my knowledge has not been disclosed before, was part of an increasingly sophisticated machinery of war that was becoming intoxicating in its efficiency.40
Because Pakistan and Yemen were sovereign nations not at war with the United States, under American law it was the CIA that was officially in charge of the drone programme. But in practice the selection of targets and the operation of drones were a joint operation with the US Air Force and JSOC. Targets were typically approved by a joint committee of different intelligence agencies, both military and civilian. One JSOC officer who witnessed these decisions said, ‘It was a joint thing and there was always a mixture of intelligence. I cannot recall any strike based on human intelligence alone; there was always a great deal of technical intelligence so you could know who it was in the compound.’ He did add, however, that there was some leeway for striking based on thinner intelligence if it was a higher-value target.
The collaboration on drone assassinations – and the central role these began to take in counterterrorism policy – convinced many CIA veterans that they were witnessing an unhealthy militarization of the agency. ‘This isn’t espionage. You know, sitting there and looking through film, picking targets for Predators is something the military used to do,’ said Robert Baer, a former senior CIA operative. ‘The CIA runs human sources, National Security Agency do intercepts, the people who do overhead [photo
graphic images] do that and the military run lethal weapons. This is something the CIA was dragged into after 9/11.’ The work with Predators and killing people with them had come to completely absorb the agency, both its talent and resources.
Not only was the CIA becoming more militarized – diverted by drones as it had been by harsh interrogations – it was also getting tied up in dealing with the immediate, tactical threat: the latest al-Qaeda operational commander or group of militants who might form the next active cell in the West.
What was missing, some argued, was ‘over-the-hill’ strategic intelligence: that is, a glimpse at more than was immediately visible, that would allow decision-makers to better understand the causes of the continuing support for militant Islam and to identify other threats. These were what MI5 used to call its ‘horizon watchers’.
A decade on from the attacks of 11 September, while bin Laden was dead and al-Qaeda, as a movement, weakened, it was hard to argue that the jihadist cause and the threat it posed to the West were diminished. Little had been done to address the causes of terrorism and anti-Western feeling generally. Extreme Islamists had established safe havens along the Afghan–Pakistan border, in Yemen, Somalia and parts of West Africa. Meanwhile, there had been no progress in ameliorating the enduring Arab–Israeli conflict, which inspired such anger across the Middle East. By the end of 2011, US troops withdrew from Iraq but without defeating the Sunni radicals they had fought for over eight years. Across the border in Syria, an uprising was under way against the regime of President Bashar al-Assad; Sunni radicals played a significant part in that revolution from the beginning.