The International Atomic Energy Agency (IAEA) assessment has cleared the way for the board of governors to end the Agency’s extraordinary investigation into accusations of Iran’s past nuclear weapons work. But a closer examination of the document reveals much more about the political role that the Agency has played in managing the Iran file.

Contrary to the supposed neutral and technical role that Director General Yukiya Amano has constantly invoked and the news media has long accepted without question, the Agency has actually been serving as prosecutor for the United States in making a case that Iran has had a nuclear weapons program.

The first signs of such an IAEA role appeared in 2008 after the Bush administration insisted that the Agency make a mysterious collection of intelligence documents on a purported Iranian nuclear weapons research program the centerpiece of its Iran inquiry. The Agency’s partisan role was fully developed, however, only after Amano took charge in late 2009. Amano got U.S. political support for the top position in 2009 because he had enthusiastically supported the Bush administration’s pressure on Mohammed ElBaradei on those documents when Amano was Japan’s permanent representative to the IAEA in 2008.

Amano delivered the Agency’s November 2011 report just when the Obama administration needed additional impetus for its campaign to line up international support for “crippling sanctions” on Iran. He continued to defend that hardline position and accuse Iran of failing to cooperate as the Obama administration sought to maximize the pressure on Iran from 2012 to 2015.

When the Obama administration’s interests shifted from pressuring Iran to ensuring that the nuclear agreement with Iran would be completed and fully implemented, Amano’s role suddenly shifted as well. In late June, according to Iranian officials involved in the Vienna negotiations, Secretary of State John Kerry reached agreement with both the Iranians and Amano that the “possible military dimensions” (PMD) issue would be resolved through a report by Amano before the end of the year.

Based on that agreement, Amano would write a report that would reach no definitive conclusion about the accusations of nuclear weapons work but nevertheless bring the PMD inquiry to an end . The report was still far from even-handed. It could not be, because Amano had embraced the intelligence documents that the United States and Israel had provided to the IAEA, around which the entire investigation had been organized.

Dodgy Intelligence Documents

Iran had insisted from the beginning that the intelligence documents given to the IAEA were fraudulent, and ElBaradei had repeatedly stated publicly from late 2005 through 2009 that the documents had not been authenticated. ElBaradei observes in his 2011 memoirs that he could never get a straight answer from the Bush administration about how the documents had been acquired. Different cover stories had been leaked to the media over the years suggesting that either an Iranian scientist involved in the alleged weapons program or a German spy had managed to get the documents out of Iran. But in 2013, former senior German foreign office official Karsten Voigt revealed to me in an interview that German intelligence had obtained the documents in 2004 from a sometime source whom they knew to be a member of the Mujahideen E-Khalq (MEK). A cult-like Iranian exile terrorist group, MEK had once carried out terror operations for the Saddam Hussein regime but later developed a patron-client relationship with Israeli intelligence.

Quite apart from the unsavory truth about the origins of the documents, the burden of proof in the IAEA inquiry should have been on the United States to make the case for their authenticity. There is a good reason why U.S. judicial rules of evidence require that “the proponent must produce evidence sufficient to support a finding that the item is what the proponent claims it is.”

But instead Amano has required Iran, in effect, to prove the negative. Since it is logically impossible for Iran to do so, that de facto demand has systematically skewed the entire IAEA investigation toward the conclusion that Iran is guilty of the covert activities charged in the intelligence documents. And the Agency has reinforced that distorted frame in its final assessment by constantly making the point that Iran possesses technology that could have been used for the development of a nuclear weapon. Every time Iran produced evidence that a technology that the IAEA had suggested was being used for the development of nuclear weapons was actually for non-nuclear applications, the Agency cast that evidence in a suspicious light by arguing that it bore some characteristics that are “consistent with” or “relevant to” work on nuclear weapons. The “final assessment” uses that same tactic to frame not only Iranian development of various technologies but its organizations, facilities, and research activities as inherently suspicious regardless of evidence provided by Iran that they were for other purposes.

Another tactic the IAEA had used in the past to attack Iran’s credibility is the suggestion that the government actually made a partial confession. In May 2008, the IAEA had claimed in a quarterly report that Iran “did not dispute that some of the information contained in the documents was factually accurate but said the events and activities concerned involved civil or conventional military applications.” That statement had clearly conveyed the impression that Iran has admitted to details about activities shown in the documents. But in fact Iran had only confirmed information that was already publicly known, such as certain names, organizations, and official addresses, as the IAEA itself acknowledged in 2011. Furthermore, Iran had also submitted a 117-page paper in which it had pointed out that “some of the organizations and individuals named in those documents were nonexistent.”

The IAEA resorted to the same kind of deceptive tactic in the final assessment’s discussion of “organizational structure.” It stated, “A significant proportion of the information available to the Agency on the existence of organizational structures was confirmed by Iran during implementation of the Road-map.” That sentence implied that Iran had acknowledged facts about the organizations that supported the purported intelligence claims of a nuclear weapons research program. But it actually meant only that Iran confirmed the same kind of publicly available information as it had in 2008.

On the issue of whether an Iranian organization to carry out nuclear-weapons research and development had existed, the final assessment again uses suggestive but ultimately meaningless language: “[B]efore the end of 2003, an organizational structure was in place in Iran suitable for the coordination of a range of activities relevant to the development of a nuclear explosive device.”

Similar language implying accusation without actually stating it directly can be found in most of the assessments in the document. In the section on “procurement activities,” the report refers to “indications of procurements and attempted procurements of items with relevance, inter alia, to the development of a nuclear device.” That language actually means nothing more than that Iranians had sought to purchase dual-use items, but it preserves the illusion that the procurement is inherently suspicious.

EBW and MIP

The use of “relevance” language was, in fact, the IAEA’s favorite tactic for obscuring the fact that it had no real evidence of nuclear weapons work. On the issue of the purported intelligence documents showing that Iran had developed and experimented with Exploding Bridge-Wire (EBW) technology for the detonation of a nuclear weapon, Iran had gone to great lengths to prove that its work on EBW technology was clearly focused on non-nuclear applications. It provided detailed information about its development of the technology, including videos of activities it had carried out, to show that for the objective of the work was to develop safer conventional explosives.

The IAEA responded by saying “ that the EBW detonators developed by Iran have characteristics relevant to a nuclear device.” By that same logic, of course, a prosecutor could name an individual as a suspect in a crime simply because his behavior showed “characteristics relevant” to that crime.

A similar tactic appears in the assessment of the “initiation of high explosives” issue. The 2011 IAEA report had recorded the intelligence passed on by the Israelis that Iran had done an experiment with a high explosives detonation technology called multipoint initiation (MIP) that the Agency said was “consistent with” a publication by a “foreign expert” who had worked in Iran. That was a reference to the Ukrainian scientist Vyacheslav Danilenko, but he was an expert on producing nanodiamonds through explosives, not on nuclear weapons development. And the open-source publication by Danilenko was not about experiments related to nuclear weapons but only about measuring shock waves from explosions using fiber optic cables.

The 2011 report also had referred to “information” from an unnamed member state that Iran had carried out the “large scale high explosives experiments” in question in the “region of Marivan.” In its final assessment, the Agency says it now believes that those experiments were carried out in a “location called ‘Marivan’,” rather than in the “region of Marivan.” But although Iran has offered repeatedly to allow the IAEA to visit Marivan to determine whether such experiments were carried out,the IAEA has refused to carry out such an inspection and has offered no explanation for its refusal.

The Agency relies on its standard evasive language to cover its climb-down from the 2011 assessment. “The Agency assesses that the MPI technology developed by Iran has characteristics relevant to a nuclear device,” it said, “as well as to a small number of alternative applications.” That wording—combined with its refusal to make any effort to check on the one specific claim of Iranian experiments at Marivan—makes it clear that the Agency knows very well that it has no real evidence of the alleged experiments but is unwilling to say so straightforwardly.

The Agency did the same thing in regard to the alleged “integration into a missile delivery system.” A key set of purported intelligence documents had shown a series of efforts to integrate a “new spherical payload” into the existing payload chamber of the Shahab-3 missile. The final assessment avoids mention of the technical errors in those studies, which were so significant that Sandia National Laboratories found through computer simulations that not a single one of the proposed redesign efforts would have worked. And it later became apparent that Iran had begun redesigning the entire missile system—including an entirely different reentry vehicle shape from the one shown in the drawings—well before the start date of the purported nuclear weapons work.

But the IAEA was only interested in whether the workshops portrayed in the purported intelligence were in fact workshops used by the Iranian government. Iran allowed the Agency to visit two of the workshops, and the final assessment declares that it has “verified that the workshops are those described in the alleged studies documentation” and that “the workshop’s features and capabilities are consistent with those described in the alleged studies documentation.”

Flawed Computer Modeling

One of the most egregious cover-ups in the assessment is its treatment of the alleged computer modeling of nuclear explosions. The agency recalled that it had “received information from Member States” that Iran had done modeling of “nuclear explosive configurations based on implosion technology.” Unfortunately for the credibility of that “information,” soon after that 2011 report was published someone leaked a graph of one of the alleged computer modeling efforts attributed to Iran to Associated Press reporter George Jahn. The graph was so similar to one published in a scholarly journal in January 2009 that Scott Kemp, an assistant professor of nuclear science and engineering at the Massachusetts Institute of Technology (MIT), said he suspected the graph had been “adapted from the open literature.”

Furthermore the information in the graph turned out to be inaccurate by four orders of magnitude. In response to that revelation, a senior IAEA official told Jahn that the Agency knew that the graph was “flawed” as soon as it had obtained it but that IAEA officials “believe it remains important as a clue to Iranian intentions.” In fact the official revealed to Jahn that the Agency had come up with a bizarre theory that Iranian scientists deliberately falsified the diagram to sell the idea to government officials of a nuclear explosion far larger than any by the United States or Russia.

That episode surely marks the apogee of the IAEA’s contorted rationalizations of the highly suspect “information” the Agency had been fed by the Israelis. In the final report, the Agency ignores that embarrassing episode and “assesses that Iran conducted computer modeling of a nuclear explosive device prior to 2004 and between 2005 and 2009,” even though it describes the modeling, enigmatically, as “incomplete and fragmentary.

The assessment further “notes some similarity between the Iranian open source publications and the studies featured in the information from Member States, in terms of textual matches, and certain dimensional and other parameters used.” Unless the Agency received the “information” from the unidentified states before the dates of the open-source publications, which one would expect to be noted if true, such similarities could be evidence of fraudulent intelligence rather than of Iranian wrong-doing. But the assessment provides no clarification of the issue.

Nuclear Material

On the issue it calls “nuclear material acquisition,” however, the Agency makes a startling retreat from its previous position that has far-reaching implications for the entire collection of intelligence documents. In its 2011 report, the IAEA had presented a one-page flow sheet showing a process for converting “yellow cake” into “green salt” (i.e., uranium that can be enriched) as a scheme to “secure a source of uranium suitable for use in an undisclosed enrichment program.”

But the final assessment explicitly rejects that conclusion, pronouncing the process design in question “technically flawed” and “of low quality in comparison with what was available to Iran as part of its declared nuclear fuel cycle.” In other words, Iran would have had no rational reason to try to seek an entirely new conversion process and then turn the project over to incompetent engineers. Those were precisely the arguments that Iran had made in 2008 to buttress its case that the documents were fabricated.

The assessment carefully avoids the obvious implication of these new findings—that the anomalies surrounding the “green salt” documents make it very likely that they have were fabricated. To acknowledge that fact would cast doubt on the entire collection. But the surprising backtracking on the “green salt’ evidence underlines just how far the IAEA has gone in the past to cover up awkward questions about the intelligence at the center of the case.

Now that the Obama administration has settled on a nuclear agreement with Iran, the IAEA will no longer have to find contorted language to discuss Iran’s past and present nuclear program. Nevertheless, the Agency remains a highly political actor, and its role in monitoring and reporting on the implementation of the agreement may bring more occasions for official assessments that reflect the political interest of the U.S.-led dominant coalition in the IAEA board of governors rather than the objective reality of the issue under review.

Gareth Porter is an independent investigative journalist and winner of the 2012 Gellhorn Prize for journalism. He is the author of the newly published Manufactured Crisis: The Untold Story of the Iran Nuclear Scare.

Counter punch,