Why did this week's National Intelligence Estimate so completely change the U.S. view on Iran's nuclear weapons program? Was there a U.S. spy inside the program? A defector? A snippet of electronically purloined communications?
According to current and former U.S. intelligence officials, the answer is not so easy. While the United States did come up with some new information—the "great discovery" President Bush described Tuesday—the estimate was the result of a subsequent review of "hundreds of pieces of data," some old, some new, as one current official put it. And not all of it was clandestinely obtained.
Some was what spies call "open source intelligence," material in the public domain.
"There was new collection as well as new analysis," said one senior U.S. intelligence official.
"There were multiple streams of intelligence that provided new information as well as illuminated old facts," said another.
The key piece of information—and the officials will not describe it further—came into U.S. hands earlier this year, perhaps as late as this summer. The former senior intelligence official told NBC News that while he wasn't "blown away" by the new intelligence—"it was not a holy shit moment" it pushed reassessment of the 2005 NIE…as did a congressional review mandated last year. That NIE stated that Iran had moved inexorably towards a nuclear weapon. It was the reassessment, not the "great discovery," that was "compelling stuff," said the official.
So compelling, said the former official, that there was "no disagreement" on the key judgment—that Iran had stopped its weapons work in fall 2003…and not just among the 16 intelligence agencies, but also among those outside the community who reviewed the draft.
Another former senior intelligence official described the new intelligence to NBC News as "something big"…big enough that it caused the intelligence community to rethink its 2005 NIE.
The reassessment took two tracks, say current and former officials. In the first, the intelligence community prioritized its intelligence assets, pushing "several streams of intelligence, several disciplines"—not just human sources or electronic eavesdropping or spy satellites. It was "an all-source" review, said a current intelligence official.
In the second, the community reassessed what it had used to develop the first NIE in 2005…a "scrub" of those original sources. In fact, some of those sources in retrospect were seen as suspect.
One former official said that while he does not know the details, he says whatever happened had to have been "recent and sourced at multiple levels," perhaps with the help of friendly—and maybe some not so friendly—intelligence services, referring to those in the Middle East and perhaps the former East Bloc.
"You don't make this dramatic a change in assessment without something so solid as to be irrefutable," he said. "It could not be one human source or one signal intelligence channel. It has to be sourced multiple ways by both to avoid falling into a disinformation trap."
A current official agreed. "I don't need to tell you that we re not going to say we believe something with ‘high confidence' based on a single stream of intelligence, no matter how good we think it is."
Moreover, the former official notes that if there was a human source critical to the new intelligence, he or she is unlikely to be still in Iran. A public disclosure like this would jeopardize their personal safety. He suggests it could also be a recent former official who had defected quietly or an Iranian official who was a U.S. asset and then died. (The intelligence community denies that Ali Reza Asgari, a Revolutionary Guard official who disappeared in Turkey this summer, was critical to the reassessment.)
Bill Harlow, formerly a member of the CIA's senior staff and now an NBC News analyst, notes: "Types of information would include human intelligence—spying, electronic intelligence—intercepted communications, satellite analysis a whole variety of information that would be used all together to form this new picture."
Harlow and other current intelligence officials also said the United States would rely on the work of friendly intelligence services—and maybe some non-friendly services. The service or services that provide such intelligence would have had to sign off on the release of the NIE, or at least could have requested anything it had provided be excised. There's no indication that took place.
Among the open source material checked were two pieces of video from 2005: news file of former Iranian President Mohammed Khatami visiting the Iranian uranium enrichment facility in Natanz in March 2005 and an Iranian government video from the same time frame laying out the Iranian nuclear power program. Although officials say that neither was critical to the overall assessment that the program had been halted, both helped them gauge the progress of the enrichment program. (The second video came in two forms: one for internal distribution which featured Natanz and another who was identical except that all references to Natanz had been excised.)
Bottom line, for many people though is why is the 2007 NIE is so different from the 2005 NIE. Charles Duelfer is a former UN weapons inspector who later reviewed the intelligence that helped push the United States in the Iraq war.
"The problem is that you get information with a delay. We learned about the decision in 2003 in 2007, said Duelfer. "It's a bit like dodging a ball, you move forward on one assumption then you get more information, it is a very difficult task to get this information and draw conclusions."
So often, say officials, you wait for something new to evolve, to challenge the previously held positions, which is exactly what happened in this case.