My first experience as a weapons inspector was in implementing the Intermediate-Range Nuclear Forces Treaty between the US and the former Soviet Union, and I’m a firm believer that on-site inspections should be part of any arms control agreement. As a United Nations weapons inspector in Iraq, I worked closely with the IAEA to investigate Iraq’s past nuclear weapons programme, and I have confidence in the IAEA’s ability to implement the Nuclear Non-Proliferation Treaty. The provisions of the NPT are at the heart of the framework agreement with Iran, and the measures contained in it – which include sophisticated remote monitoring, and environmental sampling at undeclared facilities – should be more than adequate to establish whether or not it has diverted any nuclear material to a weapons programme. The framework agreement also calls for a range of verification measures beyond those required by the NPT. These cover centrifuge production and aspects of the uranium fuel cycle such as mining and processing, and are needed to verify that Iran isn’t engaged in covert uranium enrichment using a secret cache of centrifuges and unaccounted-for stocks of uranium ore. No notice inspections to investigate ‘possible military dimensions’, however, go far beyond anything required by the NPT. The question is whether such an intrusive measure is warranted or whether, as Iran argues, the inspections would infringe its legitimate security interests.

The facts appear to support Iran’s position. Countries subjected to intrusive no notice inspections have to be confident that the process isn’t actually an intelligence-led operation aimed at undermining their legitimate interests. The nuclear framework agreement with Iran doesn’t require the IAEA to accept anything Iran declares at face value, but none of its protocols justifies no notice inspections of military sites. Iran signed the Joint Plan of Action in 2013, and has abided by the verification conditions it required without incident. This track record should count in its favour, especially when you consider the dubious results of no notice inspections since they were first carried out in 1991.

*

Until the late 1980s, on-site inspections hadn’t been included in any postwar arms control agreements. For decades, negotiators from the US and the Soviet Union discussed different verification measures, including remote sensor monitoring, overflights and ‘national technical means’ (a euphemism for spy satellites). But whenever the US raised the possibility of on-site inspections, the Soviet Union would protest, believing that teams of inspectors visiting sensitive sites would be used as a cover for intelligence-gathering. For the Americans, on-site inspections became a litmus test for judging how serious the USSR was about a particular arms control issue. In July 1987, when the Soviet Union accepted a US plan for verification of disarmament that included an intensive programme of on-site inspections, many American negotiators were taken by surprise. The Intermediate-Range Nuclear Forces Treaty (INF) was signed that December, and on-site inspections became an essential part of disarmament agreements.

By ratifying the INF treaty, the United States and the Soviet Union agreed that teams of inspectors would supervise the destruction of missiles, conduct ‘baseline’ inspections of all declared facilities and regular monitoring inspections at each country’s largest missile production facility: the Hercules Plant in Utah, and the Votkinsk Machine Building Plant in the foothills of the Urals. Provisions for short-notice ‘challenge’ inspections – which could be at any declared site and could not be refused – were agreed on and implemented without any serious disputes. Mutual fears over the ‘inspector-spy’ gaining access to sensitive military installations soon gave way to mutual respect for the professionalism of the inspectors and the inspected.

During the 13 years that on-site inspections were in force, both parties were serious about keeping to the provisions of the INF treaty. Proposals – I know of two – to expand intelligence collection by US inspectors beyond what could be observed through serendipity were immediately rejected by the CIA. This didn’t mean there wasn’t any controversy: there was a crisis, for example, over the US installation of an X-ray imaging system known as CargoScan at Votkinsk in the spring of 1990: the Soviet Union was concerned that it might damage the propellant in its non-treaty-limited missiles. But rather than allow their differences to undermine the treaty, both parties continued to refer back to its terms when seeking a solution for any problems. The INF treaty became the template for subsequent arms control and disarmament agreements, whether bilateral (such as the US-Soviet Strategic Arms Reduction Treaty, or START) or multilateral, such as the Security Council resolutions calling for the disarmament of Iraq in the aftermath of the 1991 Gulf War. With START, the INF model worked. In Iraq it didn’t.

In the INF model, all inspection procedures were spelled out in the treaty, and what was inspected was determined by data provided by the inspected party. Intelligence played a minor role: the CIA operated two ‘gateway’ facilities – one in Frankfurt and the other at Yokota Air Base in Japan – which provided support for INF inspections. This support was logistical – equipping and arranging transport for the inspection teams – and it was never the intention that CIA intelligence should alter the course of the inspections themselves. Inspections in Iraq were initially supposed to operate in the same way, with Security Council resolutions and Iraqi declarations setting the parameters for on-site inspections. But incomplete data submissions and active concealment by Iraq made the INF model hard to follow. For Unscom, the UN programme to inspect Iraqi weapons, on which I served between 1991 and 1998, the CIA set up a ‘gateway’ operation in Bahrain, with the assistance of the British, Canadian and Australian intelligence services. Intelligence support was available only to those four countries. This led to friction within the inspection teams, and concerns about American influence over what was supposed to be a UN operation.

Two senior Americans at Unscom with considerable experience in INF inspections, the director of operations and a ballistic missile chief inspector, did their best to strike a balance between the UN’s need to maintain its independence and the CIA’s sensitivities over information security. But Iraqi obstruction made it possible for the CIA to criticise both men for being too soft on the Iraqis and having an anti-American bias. In October 1991 Colin Powell, then chairman of the Joint Chiefs of Staff, demanded that they respond to the CIA’s allegations. The charges against them were refuted, and Powell dropped his inquiry, but by the summer of 1992 both men had been pushed out of Unscom.

The CIA was in a position to make demands because intelligence provided by the US played such an important role in the Unscom inspections. A pair of Iraqi defectors had provided the CIA with information about locations in Baghdad used to hide sensitive documents from the inspectors. A joint Unscom-IAEA inspection team was put together in a rush, the critical mission planning done not by the director of operations or the veteran INF chief inspector but by the CIA. The result was what’s known as the ‘parking lot incident’: in September 1991, the inspection team seized thousands of documents, including some that provided clear evidence that Iraq had an undeclared nuclear weapons programme. The team was led by an aggressive IAEA inspector called David Kay, though it was not really an IAEA operation but a US one: the deputy chief inspector, the American diplomat Bob Gallucci, called most of the shots. ‘The team,’ Gallucci said in 2001, ‘was very, very special … we had a lot of team members with special skills, especially people who knew how to search buildings.’ Gallucci recalled sitting with another inspector, who ‘looked at the fellow who was driving the vehicle, who was one of our “special people”, and he said to me: “He does not look like a physicist.” And I said: “It’s just because he has a really thick neck. Is that what you’re thinking?” And he said: “Yes, that … and the crew cut. Where did you get him?” I answered: “Well, there was an ad in the New York Times.”’ In fact, these ‘special’ team members, trained in ‘close target reconnaissance’ and ‘surreptitious entry’, worked for the Combat Applications Group and the Special Activities Division, better known as Delta Force and the CIA. And after the success of the parking lot incident the US relied on them to conduct all no notice inspections in Iraq. From the American perspective, Unscom now had a model of on-site inspection that worked.

*

I got my first taste of the realities of no notice inspections in December 1991 at a US-run briefing in an aircraft hangar in Bahrain. My notes from that day: ‘The inspection is like a raid. Surprise, speed and decisive action will carry the day.’ The instructor was a man of military bearing with a non-regulation haircut and facial hair, an expert in what he called ‘sensitive site exploitation’ – the art of rapidly entering and evaluating a room or structure for persons and materials of interest, and securing anything worthwhile. Other members of the team included a number of US paramilitary types, French Marine commandos, various British soldiers of fortune, and the odd rocket scientist, chemist, biologist and nuclear physicist. It could have been a casting call for Mad Max.

When we arrived in Iraq, our convoys of four-wheel-drive vehicles raced through city streets or across the desert, with sensor-laden helicopters and U-2 high-altitude reconnaissance aircraft above and high-resolution spy satellites providing further imagery. Later inspections included covert operators whose task was to intercept sensitive Iraqi communications, as well as networks of agents who reported on what was happening in and around the targeted areas. The parking lot incident was the template for these raid-like inspections: highly sensitive intelligence was released by the US on condition that the inspectors would protect the source and make sure they surprised their targets.

But in the summer of 1996, the CIA used paramilitaries assigned to an Unscom inspection team to assist in a failed coup attempt against Saddam Hussein – an action which Unscom had no knowledge of, and would never have permitted – and from then on Bob Gallucci’s special people were no longer made available by the US government. By this time, however, Unscom had significant experience in no notice inspections. By 1997 I had started to run a five-day Inspector Operations Course before each major inspection round. The techniques used in the raids themselves remained fundamentally unchanged, although some new tactics, such as the use of remote cameras, had been added. Team members were instructed in subjects ranging from cultural sensitivity (‘Your behaviour must be beyond reproach at all times’) to attitude (‘You are the Alpha Dog’), along with training in site exploitation, document processing and tactical convoy driving.

Inspectors’ résumés no longer listed work in places like Mogadishu, Khartoum or San Salvador, but rather involvement in Unscom missions that had often turned into intense confrontations between inspector and inspected. The change led to a new ‘inspector culture’ that was alien to all who weren’t part of the tribe. A reporter from Le Monde observed this in action at the Canal Hotel in Baghdad, which served as the headquarters of the United Nations in Iraq: ‘One lot wore jeans, knocked back cans of beer, played darts and put on deafening disco music. The other group wore ties, sipped gin and tonics, watched CNN news and tried to turn down the volume of the music.’ Inspectors were derided by their humanitarian colleagues as ‘cowboys’, and the humanitarian workers were referred to by inspectors as ‘bunny huggers’.

There’s no doubt that the Unscom cowboys had a bit of an attitude, but it sprang from unfulfilled expectations, not arrogance. Each inspection began like a cup final, only to lose its excitement because of Iraqi obstructionism, external political interference (usually from the US) or Security Council ambivalence – sometimes all three. Team morale remained high, but cynicism crept in: our theme song was U2’s ‘I Still Haven’t Found What I’m Looking For’, and every laptop had a copy of a clip from the movie Spaceballs (‘Find anything yet? We ain’t found shit!’) that was played at the end of each day, as we returned empty-handed to prepare our daily reports.

*

During my seven years as an Unscom inspector, I worked with the CIA, the Israeli Aman, the Dutch BVD and the German BND. But my closest relationship was with British intelligence. From 1991 to 1996, our dealings were managed through Operation Rockingham, a Defence Intelligence Service organisation that served as a clearing-house for all the intelligence support provided to Unscom by the UK. By 1996 most of Unscom’s leads had dried up and my need for actionable information was such that the Secret Intelligence Service (SIS) agreed to deal with me directly. The SIS assigned me a codename – Dark Knight – for use in our correspondence (Richard Butler, Unscom’s executive chairman, was Dark Prince).

The sites for Unscom inspections were originally determined by declarations made by Iraq. In the first statements it provided to the UN, in April 1991, it underestimated its holdings of chemical weapons and ballistic missiles, and failed to acknowledge either a biological or nuclear weapons programme. Unscom was forced to turn to member governments for new intelligence to make up for the information shortfall. The inspection process temporarily revived: information from a defector led to the parking lot incident, which exposed the existence of the country’s nuclear weapons programme; satellite imagery detected a still existing covert missile force; and contracts that documented the purchase of complex growth media for propagating bacteria compelled Iraq to admit it had a biological weapons programme. But even this intelligence had a ‘use by’ date. What was lacking was a source inside Iraq who could update the information provided by defectors. The CIA refused to discuss the agents it might be controlling inside Iraq and how they might be able to help Unscom. SIS was much more accommodating, especially after a meeting I attended at its headquarters in Vauxhall in August 1997. Debriefing reports coming out of the gateway office in Bahrain had highlighted the name of a Special Republican Guard officer who had had contact with the inspection team. It happened that this officer had been in contact with relatives in England, and had expressed dissatisfaction with life in Iraq. SIS had assigned him the codename Ultimate Goal, but since it no longer had a presence inside Iraq, the recruitment effort had gone nowhere.

Enter Unscom. At Vauxhall the SIS official responsible for the Middle East (I’ll call him ‘the Don’) approached me about a matter of great sensitivity. It was my inspection team that had made contact with Ultimate Goal, and I’d spent a significant amount of time questioning him about his role in concealing material from Unscom. ‘Could you arrange for another inspection of his office?’ the Don asked. I told him that I could. The Don then introduced me to an Arabic-speaking junior officer (the Junior Executive), and we hatched a plan. I would get the Junior Executive into Ultimate Goal’s office, and then create a distraction while the Junior Executive conducted a quick assessment of the situation before deciding whether or not to place in Ultimate Goal’s desk instructions on how to make contact with SIS. I ran this by my boss, Richard Butler, when I returned to New York, and to my surprise he signed off on the proposal without any debate. The next month, the Junior Executive gained access to Ultimate Goal while I kept his colleagues busy. I don’t know what the result of the mission was. ‘We won’t be able to tell you if this worked or not,’ the Don had told me. ‘What I can promise you is that if and when we get information that is of use to you and your team, you will get it.’

The continued failure of Unscom to uncover significant proscribed activities and material in Iraq, combined with the political fallout from the no notice inspections, caused Unscom’s collapse in 1998. SIS played a role in the final drama: an agent in Iraq provided information about ballistic missile components hidden in a Baath Party property in Baghdad. The site was due to be inspected in August 1998, but the mission was aborted after the Iraqis ceased all co-operation with Unscom. In December 1998 Unscom tried again to inspect it, prompting a confrontation with Iraq that led to the withdrawal of Unscom and to Operation Desert Fox, a 72-hour aerial assault by the US and the UK. Unscom inspectors never returned. In September 2002, I went back to Iraq to film a documentary about disarmament and visited the Baath Party property in question. The SIS report contained errors in critical details about its layout, bringing into question the source’s credibility; it’s unlikely anything would have been found had an inspection gone ahead.

*

Unmovic, the United Nations Monitoring, Verification and Inspection Commission, was created by the UN Security Council in December 1999. It was designed to be different from its predecessor, staffed by employees paid by, and ostensibly loyal to, the UN; Unscom had used ‘experts on mission’ loaned from its member governments. Each inspector was required to attend a month-long course of instruction; in February 2003 Unmovic’s executive chairman, Hans Blix, told attendees at one such course that there would be ‘detailed lectures about various Iraqi weapons programmes, about the result of past inspections, about the craft and tools of inspection, the rights and duties of inspectors in Iraq and about the history, culture and religions of Iraq’. An Unmovic inspector, he said, should be ‘driving and dynamic – but not angry and aggressive’; ‘ingenious – but not deceptive’; ‘keeping some distance – but not arrogant or pompous’. Between its creation and the return of inspectors to Iraq in November 2002 Unmovic had nearly three years to prepare. Once on the ground, it conducted 750 inspections at 550 sites. Most of them were routine, familiarising Unmovic inspectors to sites already inspected by Unscom. But there were also no notice inspections of sites that hadn’t been inspected before, based on intelligence provided by supporting governments. The vast majority of these inspections produced no results: the intelligence was either wrong or out of date. But on one occasion it was dead-on: the inspection of the home of Fahel Hassan Hamza, a scientist who in the 1980s had conducted work related to the laser separation of radioisotopes. A cache of three thousand documents was discovered, most of which related directly to Hamza’s work with lasers. It looked as if the Unmovic model had succeeded where Unscom had failed.

Inspectors have remained tight-lipped about the tip that led to the inspection of Hamza’s house. The British government’s ‘Review of Intelligence on Weapons of Mass Destruction’ (better known as the Butler Review), published in July 2004, attributes the intelligence to the UK, and most likely to SIS. According to the Butler Review, SIS had six agents in Iraq at the time. (I can’t ascertain if one of them was Ultimate Goal.) The nuclear-related reporting appeared to come from two of these sources, both termed ‘new’, neither of whom had direct experience in current WMD programmes. The British were more reticent about sharing human intelligence sources with Unmovic than they had been with Unscom. Unmovic’s new ‘independent’ profile meant it was willing only to receive information.

Unmovic’s point of contact for receiving foreign intelligence, the Canadian ex-intelligence officer Jim Corcoran, was cleared to handle sensitive information, but SIS was less confident about the rest of the Unmovic staff or its procedures for transmitting sensitive data into Iraq. When Corcoran met with SIS, they insisted that intelligence had to be carried into Iraq by hand, and that knowledge of each report had to be limited to as few people as possible. Two inspectors – Kay Mereish, a retired US colonel who had worked at the Medical Research Institute of Infectious Diseases in Maryland, and Martin Fosbrook, a British biologist – flew back to New York so that Corcoran could brief them, along with Dimitri Perricos, a veteran IAEA inspector who served as chief inspector for this mission, in a secure space, precluding any need for conversation in Iraq of a sort that shouldn’t be overheard. On 14 January the three inspectors returned to Baghdad. Two days later they inspected Faleh Hamza’s house.

On the morning of 16 January 2003 a convoy of white UN vehicles left the Canal Hotel, accompanied by their Iraqi minders in a hotchpotch of civilian vehicles. The convoy crossed the Tigris and arrived in Ghazaliyah, a neighbourhood in west Baghdad, just after nine in the morning. As well as Hamza’s house, the inspectors raided the house of his next-door neighbour, Shakir al-Jibouri, another Iraqi nuclear scientist. Both men were at work, and only their wives and children were at home. The inspectors waited outside for hours while their Iraqi minders tracked down the two scientists and brought them home. Then the inspections began. The documents were found in a wooden box in a cupboard upstairs in Hamza’s house, and Perricos ordered Mereish to take them into Unmovic custody. The Iraqi government protested and a compromise was struck: Hamza would accompany the documents to the Canal Hotel, where they would be copied by the inspectors in his presence, and he would receive a complete copy. The process took hours, and Hamza claims that at one point he was separated from his Iraqi minder and approached by a female Unmovic inspector who offered to take him and his wife out of Iraq so that he could talk to the inspectors without fear of reprisal. Hamza refused, and later complained to the press about the inspectors’ ‘mafia tactics’.

Blix used the seized documents to remonstrate with the Iraqis; he said that they represented ‘a sign that not everything has been declared’. Colin Powell, then secretary of state, cited the documents as ‘dramatic confirmation’ that Saddam was concealing evidence and not co-operating with the inspections. Unnamed Western diplomats went further, and said the documents showed there was ‘ongoing work taking place in Iraq to develop nuclear weapons’. The Iraqi government publicly criticised Unmovic for inspecting a private residence, called the seized documents ‘private papers’, and claimed that their contents were already known to the inspectors, and had nothing to do with the Iraqi nuclear programme. On 14 February Mohammed ElBaradei, then the director general of the IAEA, said that the Hamza documents ‘provided some additional details about Iraq’s laser enrichment development efforts’, but ‘refer to activities or sites already known to the IAEA and appear to be the personal files of the scientist in whose home they were found. Nothing contained in the documents alters the conclusions previously drawn by the IAEA concerning the extent of Iraq’s laser enrichment programme.’ In short, the Unmovic version of the no notice inspection accomplished nothing of significance but contributed to an already confused story. On the eve of an American-led war that used Iraqi weapons of mass destruction as its raison d’être, the results of the inspections proved disastrous.

*

The history of no notice inspections in Iraq does not bode well for their use in Iran. Such inspections are intelligence-based exercises. The bulk of the intelligence underpinning the US concerns over ‘possible military dimensions’ comes from the ‘alleged studies’ documents – a series of files the IAEA obtained in 2008 which appear to show that Iran had conducted some nuclear weapons development in 2002 and 2003. Their credibility has often been called into question and the Iranians declare they are fake. There’s good cause, too, to believe that much of the remaining intelligence buttressing the CIA’s case against Iran is flawed. The strange tale of the Iranian physicist Shahram Amiri, whose defection the CIA facilitated in the spring of 2009, serves as a case in point. Amiri was for several years before his defection an American agent-in-place whose reporting was used by the CIA in formulating its assessments on Iran. But his re-defection to Iran in 2010 suggests that he may have been a double agent, calling into question all his reporting to the CIA, before and after his defection. Operation Merlin, in which the CIA attempted to pass on to Iran flawed designs for a nuclear weapon, further undermines the CIA’s credibility as a source of information about an alleged Iranian nuclear weapons programme.

If they were permitted, where would no notice inspections in Iran take place? There are two sites that the IAEA has publicly declared to be of interest. The first is Parchin, a military facility associated with the Iranian Revolutionary Guard Command. The IAEA was granted a ‘managed access’ inspection of the facility in 2005, and found nothing. In 2007, the IAEA claimed to have received new information linking Parchin to a test of a neutron initiator, the device which starts fission in a nuclear warhead, and asked to visit the site again. Iran has refused on the grounds that the basis for such an inspection is flawed. Robert Kelley, a former IAEA inspector, agreed: ‘The allegations that Iran carried out hydrodynamic experiments related to nuclear explosives in a large steel containment vessel [at Parchin] have questionable technical credibility.’ Parchin is a sensitive military facility, and Iran fears that giving inspectors access would lead to an intelligence-driven fishing expedition. The other site of interest is in Marivan, where the IAEA contends that Iran conducted large-scale explosive tests of a multi-point initiation system, which is used to initiate nuclear fission, and in doing so to activate the neutron initiator, in a weapon. The source of this allegation appears to be what Iran justifiably claims is a set of forged documents. In 2014, Iran offered to let the IAEA conduct another ‘managed access’ on-site inspection of Marivan; the IAEA declined.

‘You can’t hang your hat on a single issue,’ Garry Dillon, the former head of the IAEA’s Action Team in Iraq, told me in October 1998. ‘By insisting on investigating every minor discrepancy, regardless of the bigger picture, you’re putting process ahead of substance. In the end, all you’ll be doing is chasing ghosts.’ He was right. In Iraq, the inspection process became a vehicle for creating confrontations that undermined international confidence in Baghdad’s willingness to abide by its disarmament obligation. When Iraq finally told the truth about its weapons programmes, no one believed it. We used to joke about how often we came back from an inspection empty-handed, repeating the saying that insanity is doing the same thing over and over again and expecting a different result.

The intelligence about the ‘possible military dimensions’ of Iran’s nuclear programme is of questionable provenance and most of it is more than a dozen years old. The consequences of failure to reach a nuclear accord with Iran today are too serious for the world to embrace a process that has been so controversial while having so little impact on legitimate disarmament. This is especially true when the inspected party, as is the case with Iran, has agreed to implement stringent verification measures and has a proven track record of abiding by them. Iran has been put in the impossible position of having to prove a negative. If it accepts inspections based on allegations it knows to be baseless, then it’s opening itself up to an endless cycle of foreign intrusion into its military and security infrastructure, and the inability of inspectors to discover something of relevance will only reinforce the belief that something is being hidden. We saw this happen before in Iraq, and the end result was a war based on flawed intelligence and baseless accusations that left many thousands dead and a region in turmoil.