The analysts came to these conclusions on their own—with no interference from the White House, the State Department, or the CIA leadership, including me. In fact, all of these judgments were coordinated across the intelligence community, making them IC judgments, not just CIA ones (so if there was a conspiracy, it was a big one, involving multiple analysts and agencies). Contrary to statements by the media and a few senators, I played no role in the judgment that there had been a protest.
It is important to note that the analysts’ view was fully supported by my boss, Director Petraeus. At an NSC principals meeting the day after the attack, Petraeus outlined the analysts’ view that the attack had evolved spontaneously from a protest. Some of the principals, including Defense Secretary Panetta, pushed back, arguing that demonstrators do not show up at a protest with weapons. Petraeus defended the analysts’ work, arguing that there were so many weapons in Libya that the analysts’ judgment was indeed quite plausible.
It is true that, after all the relevant information became available, the protest judgment turned out to be inaccurate. It turned out that there had been no protest immediately outside the TMF—although some in the intelligence community believe that there was a protest nearby, and others believe that the gathering of the attackers outside the TMF just before the assault could have been interpreted by some on the scene as a protest. But the other initial judgments of the analysts have held up over time.
And the analysts did not just make up the judgment about the protest. Two things led them to that conclusion. First, a dozen or so reports—both intelligence reporting and press reporting—said there had been a protest ongoing at the time of the attack. And second, not a single piece of information in the analysts’ possession at the time they wrote the piece that was published on September 13 said there had not been a protest.
CIA’s analysts have been criticized for not reaching out to the officers who were on the ground that night at the TMF and asking them what happened, asking them if there had been a protest. But that is simply not how intelligence analysts operate. They are analysts, not investigators. They wait for information to come to them; they do not go out and gather it. Additionally, the FBI had just opened an investigation into the deaths of the four Americans, and the Bureau would have been extremely concerned if CIA officers had interviewed the witnesses to a crime before the Bureau did.
I do think that the analysts can be criticized—and therefore the Agency and I can be criticized—for not pushing those in the field harder for more and better information faster. For example, it took the FBI a number of days to write and disseminate intelligence reports from its interviews of the eyewitnesses. We should have pushed hard to get those reports much earlier.
On Friday morning, September 14, my boss David Petraeus led a team to Capitol Hill to brief the House Permanent Select Committee on Intelligence (HPSCI). He had done a similar briefing the day before at the Senate Intelligence Committee. The talking points that had been prepared for him for these two briefings paralleled what the analysts had written on the thirteenth.
I didn’t accompany the director to the HPSCI, and learned what had transpired only late that afternoon. As I was standing in the director’s conference room between two regular but important meetings—the director’s thrice-weekly update on counterterrorism and his weekly update on Syria—Director Petraeus’s chief of staff handed me a copy of talking points on Benghazi.
He said he was concerned that I was not yet aware of an important issue and that I needed to be brought into the loop—that at the morning HPSCI briefing, the ranking member of the committee, Representative C. A. “Dutch” Ruppersberger, had asked for unclassified talking points that he and others might use that coming weekend should they be asked by the media about the attacks in Benghazi. He added that Director Petraeus had agreed to the request and that a draft of the points was already circulating both inside and outside CIA. He said, “These are the talking points as they now stand.”
I learned later that the talking points had been drafted by the head of the Counterterrorism Center’s Office of Terrorism Analysis (D/OTA), who had been with Petraeus on the Hill. She had produced a draft quickly after returning to headquarters. She had coordinated this draft with substantive experts on both the analytic and operational sides of the Agency and, because of the issues associated with speaking publicly about an ongoing FBI investigation, with attorneys from our Office of General Counsel.
After she made changes that were suggested by substantive experts and by the Office of General Counsel, the D/OTA sent the draft of the talking points to CIA’s Office of Congressional Affairs (OCA), which then took the unusual step of holding a coordination session with officers from CIA’s Office of Public Affairs (OPA). No substantive experts were involved in this process.
This was a significant mistake. The OCA and OPA staffers went well beyond their expertise and responsibilities in editing the points. The officers in these two staffs made a number of changes to the draft, including changing attacks to demonstrations in the first sentence of the D/OTA draft, which had originally read “The attacks in Benghazi were spontaneously inspired by the protests at the U.S. Embassy in Cairo and evolved into a direct assault against the U.S. Consulate.…” Participants in the editing session say they do not have a clear recollection as to why they made this change, but some have said that they believed the sentence to be illogical as written: saying that “attacks” evolved into an “assault” does not make sense, because attack and assault are synonymous. In my view, the most important point here is that the concept of an attack/assault still existed in the first sentence even after this change. Again, contrary to some media and Congressional allegations, I did not make this change. In fact, it occurred before I was even aware that the talking points had been requested.
The group of Public Affairs and Congressional Affairs officials also deleted the phrase “with ties to al Qa‘ida.” They say they did so to ensure that they would not compromise the FBI investigation by prematurely attributing responsibility for the attacks to any one person or group. They had reason to be concerned about this. One of the internal CIA e-mails sent that day came from our general counsel, Stephen Preston. It said, “Folks, I know there is a hurry to get this out but we need to hold it long enough to ascertain whether providing it conflicts with express instructions from NSS/DOJ/FBI that, in light of the criminal investigation, we are not to generate statements as to who did this etc., even internally not to mention for public release.” Again, this change took place before I was aware that the talking points had been requested, which, of course, undercuts yet another of the claims about me—that I was the one to remove the reference to al Qa‘ida from the talking points. I did not do so.
I do believe that the removal of the “with ties to al Qa‘ida” language was a mistake. It did not attribute responsibility to a particular group or particular individuals in a way that would have put the FBI investigation at risk. Those words would have made the talking points better.
The OCA/OPA version was then shown to Director Petraeus, who asked for a significant addition. The director asked that language be added regarding CIA’s assessments starting months earlier regarding the deteriorating security situation in eastern Libya, as well as the warnings sent out just days before the 9/11 anniversary. Having made these changes, the Office of Public Affairs circulated the draft talking points to its counterparts around government—the State Department, NSC, FBI, National Counterterrorism Center, and others. More changes were suggested.
This was another mistake on the part of the OCA and OPA. They had no business taking the lead in coordinating the points with the rest of the government. The substantive experts in the Office of Terrorism Analysis should have been the lead. Those experts did not even realize their points were circulating among the other national security agencies.
One of the most significant changes suggested at this point was proposed by the FBI, which requested that the phrase “We do kn
ow Islamic extremists participated in the violent demonstrations” be amended to “there are indications that Islamic extremists participated in the violent demonstrations.” The FBI did not want the talking points to be so definitive in describing the perpetrators, since the investigation was just getting under way. Finally, the State Department wanted to remove an entire sentence that linked the Islamic extremist group Ansar al-Sharia to the attack—because, it reasoned, the only unclassified evidence we had that they were involved was an initial public claim by Ansar al-Sharia taking credit for the attack that had been quickly retracted by the group. All our other evidence indicating the group’s involvement was still classified at that time.
All of this occurred before I first learned of and read a draft of the talking points on that Friday afternoon standing in the director’s conference room. As I skimmed the talking points, with the director’s chief of staff standing there, one thing leaped out at me—the inclusion of the prior-warning language. While they were factually accurate, I thought that including those sentences was ill-advised and I made my views clear to the chief of staff. To begin with, the request had been to give members of Congress language they could use to describe what had happened on September 11, 2012. What CIA had done in the months, weeks, and days leading up to the attack was simply not relevant to the request. More important, I saw the language as an attempt by the Agency to thump its chest, to say, “We did our job,” and to deflect any blame from CIA to elsewhere. I thought we would pay a price for this in the relationships that make up the interagency process in Washington. Contrary to what some of the critics have said, I did not take this position to protect the State Department. I did so to protect the Central Intelligence Agency. And I made this decision well before I even knew that the State Department did not like the warning language—in direct contradiction to what several members of the House Intelligence Committee have implied in questioning my integrity in an “Additional Views” section of its report, released in late 2014.
While he never said a word as I vented about the warning language, the chief of staff’s body language suggested to me that he agreed. In fact, I believe that this is why he’d brought the talking points to my attention in the first place. I believe he thought that I would react exactly the way I did.
In addition to protecting the Agency, I also believed it was unfair to the State Department for us to say that we had warned them, without giving the department an opportunity to say what it had or had not done with those warnings. There would be plenty of time for that discussion to take place. Months later it would become clear that the State Department had not taken adequate steps to protect itself in light of our warnings in the months and weeks leading up to the 9/11 anniversary, but during that second week of September CIA had no way to know that, and I believed it would be unfair to suggest it simply to protect ourselves.
Again, a few members of the House Intelligence Committee have argued that I acted outside my purview when I removed the warning language. Since these were “CIA talking points,” such an argument is absurd. But it is particularly silly given that the primary reason I excised the material was to protect the Central Intelligence Agency.
What I didn’t know at the time was that the warning language had been inserted at the suggestion of my boss, David Petraeus. The director’s chief of staff did not tell me that. Had I known it, I would have walked into the director’s office and discussed it with him that evening. Even though I made it clear that I did not like the warning language, I made no changes to the talking points on Friday evening—this is an inaccuracy in the Senate Intelligence Committee report on Benghazi, which said I did make a change on Friday—as I told my staff I would look at the talking points after they had been fully coordinated in the interagency review process.
The next morning, Saturday, started with my executive assistant’s informing me of two things. First, the State Department, at the working level, had informed us that it objected to the inclusion of the warning language and, because of this, the talking points were in limbo (this was the first time I became aware that the department did not like the talking points). And second, Denis McDonough, Obama’s deputy national security advisor, wanted to discuss the talking points at the deputies meeting scheduled for that morning, which suggested to me that he had been made aware of State’s concern about the warning language. I mentioned all of this to Director Petraeus and his chief of staff on Saturday morning, telling them that I agreed with the State Department’s position and explaining why. Petraeus didn’t argue the point and didn’t tell me he was the one who’d asked for the language to be inserted in the first place.
By the very the end of the deputies meeting, McDonough had not raised the talking point issue, so I did. I told my colleagues that I had some concerns about the talking points and that I knew other agencies did as well. I did not say what my concerns were. I concluded by saying I would edit the talking points myself and share them with the relevant deputies before sending them to the Hill. McDonough simply said, “Thank you, Michael.”
That Saturday was “Family Day” at CIA—an annual event at which the kin of Agency employees are invited to tour our headquarters complex. Because of the nature of intelligence work, the close relatives of CIA officers are rarely allowed to visit Langley. But once per year, on Family Day, employees can bring loved ones in to view exhibits, try on disguises, look at spy gear, take a polygraph test, and tour the director’s and deputy director’s offices. So while hundreds of folks trooped through my office to say hello, I was thinking about the talking points that were waiting for me on my assistant’s desk. I finally sat down with them late that morning. While I did some significant editorial work, my main substantive contribution was to remove the warning language.
I also took out the word “Islamic” in front of “extremists,” an action for which I have also been criticized (for allegedly trying to downplay the role of al Qa‘ida in the attacks). I removed the word “Islamic” for risk mitigation. Demonstrations were occurring in many countries throughout the Muslim world because of the YouTube video defaming the Prophet Muhammad, and the last thing I wanted was to encourage any American to say anything that could make the situation worse. And I thought, incorrectly, that “extremists” would carry the same message as “Islamic extremists”—that this had been a terrorist attack.
When I coordinated the talking points with my deputies colleagues, no significant changes were made. Throughout the entire process, the White House suggested only three changes and all of them were editorial—not a single one involved an analytic judgment—undercutting the conspiracy theory that the White House played a large role in editing the talking points. Finally, before having our Office of Congressional Affairs send the talking points to the Hill, I asked that the substantive experts and Director Petraeus review and sign off on them. All did so.
One of the narratives in the media has been that I “overruled” my boss on the question of whether or not to include the warning language. Believe me, there was no overruling Director Petraeus on anything. I had a conversation with the director about the warning language, in which he did not oppose my decision to remove it, and he had the opportunity at the end of the process to ask that the warning language be added back in. He did not do so.
While I am the first to admit that the talking points could have been better—they could have been more clearly written and they could have been more robust—the analytic judgments in them were fully consistent with what CIA had written for policy-makers on the morning of the thirteenth. This included the language about the assault on the TMF having evolved from a spontaneous protest. In short, what we were allowing HPSCI members to say publicly was exactly what we had said in our classified publications. Also, and importantly, CIA did not know that the talking points would be used publicly the next day by a senior administration official. We did not know that Susan Rice was going to use them on the Sunday-morning news shows.
It was only much later�
�in the spring of 2014—that it became clear to me how UN ambassador Susan Rice had come to receive our talking points. They were embedded in a much longer set of White House–produced talking points designed to prepare Ambassador Rice to appear on the Sunday shows the next day. Again, nothing in the CIA talking points was markedly different from the finished intelligence that Rice and other senior officials had been seeing over the previous four days.
But there was something different in the White House–produced points sent to Rice’s staff. There was a phrase in the “Goals” section: “To underscore that these protests are rooted in an Internet video, and not a broader failure of policy.” The White House has argued that its talking points were not about Benghazi but about the broader protests taking place in the region. But that explanation does not hold water—because just one bullet point later in the “Goals” section of the White House talking points is the following: “To show that we will be resolute in bringing people who harm Americans to justice”—and the only place Americans had been harmed during that period was in Benghazi. My reading of the White House talking points is that they were blaming the Benghazi attack on the video—which is not something CIA did in its talking points or in its classified analysis.
The White House view that its talking points were not about Benghazi had an important consequence. That view meant that the White House talking points did not need to be publicly released in the spring of 2013 along with the other materials related to the executive branch’s public narrative on the Benghazi attacks. This put the entire focus on the CIA talking points.
The Great War of Our Time: The CIA's Fight Against Terrorism--From Al Qa'ida to ISIS Page 23
