In another thread, Bob Schacht wrote,

BTW, according to CNN, Haynes and Mukasey are claiming that “half” of what we “know” about Al Qaeda came from torture sessions. Did they really write that, and if so, I’m wondering if its puffery or true.

Using the May 30, 2005 Bradbury memo, I think I can show where it comes from–and show why it’s a totally useless claim.

Bradbury Needed to Appeal to Efficacy to Claim These Techniques Didn’t Violate the Convention Against Torture

In the May 30, 2005 Memo, Steven Bradbury spends four pages recording the effectiveness of enhanced interrogation. He does this, at least partially, to make sure he can claim that the techniques at issue don’t "shock the conscience" and therefore don’t violate the Fifth Amendment (and therefore don’t violate CAT, which is the whole point of this memo) . In particular, Bradbury resorts to efficacy when trying to distinguish between torture condemned by the State Department and that practiced by the US. Speaking of torture practiced by other countries, Bradbury claims it simply doesn’t serve the same purpose as our torture.

There is no indication that techniques are used only as necessary to protect against grave terrorist threats or for any similarly vital government interests.

And when Bradbury is trying to distinguish enhanced interrogation from SERE, he again appeals to efficacy and necessity.

… the interrogation program we consider here furthers the paramount interest of the United States in the security of the Nation more immediately and directly than SERE training.

[snip]

It follows that use of these techniques will not shock the conscience in at least some circumstances. We believe that such circumstances exist here, where the techniques are used against unlawful combatants who deliberately and secretly attack civlians in an untraditional armed conflict in which intelligence is difficult or impossible to collect by other means and is essential to the protection of the United States and its interests, where the techniques are used only when necessary and only in the interrogations of key terrorist leaders reasonably thought to have acionable intelligence, and where every effort is made to minimize unnecessary suffering and to avoid inflicting significant or lasting harm.

It bears noting that this rant goes far beyond what Bradbury elsewhere carefully laid out as the premise of his memo. But both this claim and the one dismissing State Department concerns about torture rely on his argument that the program was necessary to protect the US. 

So to accomplish his apparent task–which was to find a way to declare the CIA interrogation program did not violate CAT (after CIA’s own IG had already concluded it did), Bradbury needed to be able to say it was effective.

The CIA IG Report Appears to Have Been Ambivalent about the Value of Enhanced Interrogation Program

Now, in addition to the overriding problem presented by the IG Report’s conclusion that the CIA interrogation program violated the Convention Against Torture, the IG Report presented another problem for Bradbury. Based on the citations that appear in this memo, the IG Report appears to have been very ambivalent about the value of the enhanced interrogation program. For example, the IG Report notes it is difficult to measure the efficacy of the interrogations.

As the IG Report notes, it is difficult to determine conclusively whether interrogations have provided information critical to interdicting specific imminent attacks. See id. at 88. And, because the CIA has used enhanced techniques sparingly, "there is limited data on which to assess their individual effectiveness." Id at 89.

(There’s a third reference to the IG Report just below these two, but it appears to be a somewhat gratuitous reference on Bradbury’s part so he could pretend he had addressed all the IG Report’s concerns.)

In addition, the IG Report notes that the value of the program cannot be measured by individual pieces of information.

According to the CIA Inspector General:

CTC frequently uses the information from one detainee, as well as other sources, to vet the information from another detainee. Although lower-level detainees provide less information than the high value detainees, information from these detainees has, on many occasions, supplied the information needed to probe the high value detainees further. … [T]he triangulation of intelligence provides a fuller knowledge of Al-Qa’ida activities than would be possible from a single detainee.

IG Report at 86. As illustrated below, we understand that even interrogations of lower-tier high value detainees supply information that the CIA uses to validate and assess information elicited in other interrogations and through other methods.

Now, I can’t be sure without the full context of the IG Report, but Bradbury appears to be pulling a rhetorical switch here. The IG Report passage appears to say that CIA should not focus exclusively on these high value individuals, because without the information from "lower level detainees" (note, not "lower-tier high value detainees," which Bradbury uses but which in this memo has no real meaning), interrogators can’t really get information from high value detainees anyway. This "triangulation" approach actually would seem to support the FBI method of interrogation more than the CIA method. But, in a discussion about the efficacy of enhanced techniques purportedly used only with high value detainees, Bradbury promptly elides the difference between interrogation of other detainees with that of high value detainees, and in so doing ascribes some of the value of more mundane interrogation to the enhanced interrogation program.  

Now, Bradbury calls his discussion of these two points "caveats," suggesting that he reluctantly cedes the IG Report’s doubts about the enhanced interrogation methods. 

He only makes one reference to the IG Report in a manner that appears to be positive–but Bradbury claims it supports a point it doesn’t appear to. Here’s the context:

In particular, the CIA believes that it would have been unable to obtain critical information from numerous detainees, including KSM and Abu Zubaydah, without these enhanced techniques. Both KSM and Zubaydah had "expressed their belief that the general US population was ‘weak,’ lacked resilience, and would be unable to ‘do what was necessary’ to prevent the terrorists from succeeding in their goals." [Effectiveness Memo] at 1. Indeed, before the CIA used enhanced techniques in its interrogations of KSM, KSM resisted giving any answers to questions about future attacks, simply noting, "Soon, you will know."  Id. We understand that the use of enhanced techniques in the interrogation of KSM, Zubaydah, and others, by contrast, has yielded critical information. See IG Report at 86, 90-91 (describing increase in intelligence reports attributable to use of enhanced techniques). 

Note what he’s doing here. In support of his claim that "enhanced techniques … yielded critical information" he provides citations that prove enhanced techniques yielded a larger number of reports. Bradbury here is conflating quality–"critical"–with sheer number in a bid to pretend the IG Report said something that–at least from his citations of it–it didn’t say.

In other words, faced with the need to use efficacy as a way to claim the interrogation program didn’t violate CAT, Bradbury was also faced with the IG Report’s assessments which–at best, at least according to his citations–only accepts that the enhanced interrogation methods led to an increase in the number of intelligence reports. 

CIA Produced Two Documents in Preparation for this Memo

So Bradbury got the CIA to create two documents for him that argue for the efficacy of the program.

Bradbury primarily cites two documents to make his claim that the interrogation program was effective (along with some older intelligence reports):

• Memorandum for Steven G. Bradbury, Principal Deputy Assistant Attorney General, Office of Legal Counsel, from [redacted], DCI Counterterrorist Center, Re: Effectiveness of the CIA Counterintelligence Interrogation Techniques (March 2, 2005) ["Effectiveness Memo"]
• Fax from [redacted], DCI Counterterrorist Center, Briefing Notes on the Value of Detainee Reporting (April 15, 2005) ["Briefing Notes"]

The descriptions make clear that both these documents were created for him. And, both these documents were created in the months leading up to this memo. In other words, these documents appear to have been created precisely to give Bradbury what he needed–the ability to argue the program was effective.

And largely based on these two documents, Bradbury provides a page and a half of specific intelligence derived from enhanced interrogation of KSM and Abu Zubaydah. Even in that page and a half, there are factual problems with Bradbury’s description. There’s information included that we know was available prior to their detention; there’s information included that was reportedly collected through persuasive rather than coercive interrogation; much of it relates to and came from lower level detainees; it includes the Padilla dirty bomber claim.

In addition, for both KSM and AZ, Bradbury includes a footnote saying "we discuss only a small fraction of the intelligence CIA interrogators have obtained" from them. Perhaps there is more important intelligence they got. Perhaps Bradbury has simply left out the inaccurate information. But we don’t know because he doesn’t give it to us.

And to be fair, there’s a better part of a page listing the intelligence gained from AZ that is entirely redacted.

(Also note, I can’t be sure, because of the huge redaction that appears in this section, but Bradbury doesn’t appear to deal with intelligence gotten from Rahim al-Nashiri, which I find a notable omission since waterboarding was supposed to have been immediately successful with him.)

From a qualitative standpoint, what Bradbury has given us does not make a convincing case for the importance of the intelligence gathered through torture. More importantly, Bradbury stops far short of providing an assessment of all the intelligence gained through torture, to weigh what was valid and important against the intelligence that turned out to be useless. So while Bradbury’s qualitative argument that enhanced interrogation is unconvincing, he doesn’t even try to address the claim that torture produces a lot of worthless intelligence.

The "Half the Intelligence" Claim Refers Only to Number of Reports

Which brings us, finally, to the claim the torture apologists are relying on–that half the intelligence they’ve gotten, or "half of what we know" comes from enhanced interrogation.

That claim comes–at least partly–from the Briefing Notes created for Bradbury so he’d have some way to make a claim that the torture program was effective. After all his list of specific intelligence gleaned through the program, Bradbury includes this:

More generally, the CIA has informed us that, since March 2002, the intelligence derived from CIA detainees has resulted in more than 6,000 intelligence reports and, in 2004, accounted for approximately half of CTC’s reporting on al Qaeda. See Briefing Notes at 1; see also IG Report at 86 (noting that from September 11, 2001, through April 2003, the CIA "produced over 3,000 intelligence reports from" a few high value detainees). You have informed us that the substantial majority of this intelligence has come from detainees subject to enhanced interrogation techniques. 

So this is how (or at least one possible way how) torture apologists get to half of "what we know." In a document created for this purpose, the CIA says that CIA detainees account for half of CTC’s reporting on al Qaeda. Which, I assume, doesn’t count what FBI knows about al Qaeda, for example. And that’s not all from torture–though "the substantial majority" is.

Or, you can use a different measure and realize that if a few high value detainees led to the production of 3,000 intelligence reports and two years later half of what CTC knew amounted to 6,000 reports, then those 3,000 reports from eighteen months of torture is awfully impressive!

But you’d still be talking only about the sheer number of reports!! 

Golly. The CIA is very impressed with itself that–after waterboarding KSM and AZ a combined 266 times, they spat up 3,000 intelligence reports, many of them utterly useless.

And that’s what–at least from the evidence in this OLC memo–the torture apologists’ claim appears to amount to. Self-congratulation that they’ve crafted a system designed to churn out as many intelligence reports as they can possibly churn out.

But even the torture apologists are not making a claim for the quality of that intelligence.