The SSCI Report and Its Critics: Torturing Efficacy

Polarization surrounding the SSCI Report (see here for Lawfare’s coverage) has been most pronounced on the efficacy of enhanced interrogation techniques (EITs). The Report and its supporters have proclaimed that EITs never produce useful information. Unfortunately, that pat assertion undermines the possibility of a consensus on future interrogation tactics, including a consensus that rules out coercion. If we believe that coercive treatment of detainees is wrong, we should acknowledge the trade-offs that this judgment entails, including the possible loss of some helpful information. Unfortunately, both the SSCI Report and the minority report critiquing it are more focused on debating points than on substance. Ironically, the CIA’s own response to the SSCI may come closest to a balanced discussion of the real stakes in this debate. The treatment of Abu Zubaydah has been a centerpiece of debate. The SSCI Report and its critics disagree on whether the enhanced interrogation techniques (EITs), including waterboarding, used on Zubaydah yielded useful information. (I’ll assume, as Lindsey Graham [p. 65] has stated, that waterboarding is unlawful; on whether it’s torture, see this paper by Mike Lewis.) The actual answer to the Zubaydah efficacy question is mixed. The CIA notes (p. 11) that Zubaydah revealed the existence of Jose Padilla, a US citizen who was eventually convicted of terrorism-related offenses after being detained for over three years without charges. Zubaydah was not being subjected to EITs per se at this time; however, the CIA notes, he was subjected to sleep deprivation, which was eventually approved as one of the enhanced techniques. The SSCI minority report suggests rightly that Zubaydah revealed a great deal of information during this period, which the majority report fails to fully acknowledge because of its definitional sleight of hand. The SSCI minority also rightly argues that the majority report undervalues the importance of corroboration in intelligence-gathering. Corroboration is important in a fluid investigation where multiple leads complete for a place on counterterrorism officials’ radar screen. Corroboration can help an official prioritize certain information, turning that information into actionable intelligence. Actionable intelligence was eagerly sought in the months after 9/11, when we still knew little about Al Qaeda and each day brought fresh fears of a new onslaught. Zubaydah’s flagging of Padilla may have allowed analysts to connect the dots, when other information was isolated and less susceptible to analysis. Viewed in this light, the SSCI Report may have been too facile in asserting that Pakistani officials had already informed the US about their concerns about Padilla or that US officials were already “on alert” because of “information indicating that [Padilla’s] travel was suspicious” (SSCI p. 30 n. 122). Learning that travel is “suspicious” in some generic sense is far less actionable than specific corroboration of an individual’s terrorist ties from a subject like Zubaydah, who was in a position to know. Conceding the minority’s point, we are still left with the reality that Zubaydah gave up little of further value when he was waterboarded in August of 2002. Indeed, the CIA does not dispute that Zubaydah was waterboarded and subjected to other EITs for at least two weeks after he had revealed everything he knew. That’s another risk of EIT-centered interrogations. Suppose a subject has cooperated after EITs, but some officials feel that the subject has more to offer. Further EITs may be futile, because the subject has in fact provided all the information known to him. But, as the saying goes, to a man with a hammer, everything looks like a nail. Interrogators may order another round of EITs, only to discover that the last round secured no new data. Even accepting that efficacy is an acceptable metric, the risk of EITs’ needless use toward an interrogation’s conclusion should give supporters pause. If the SSCI report was not sufficiently candid on the role of some EITs in Zubaydah’s April 2002 disclosures to the FBI, the minority’s use of statistics is even more problematic. At p. 21 of the minority report, the minority concludes that the decision not to use EITs on some detainees had a “failure rate” of over 40%. However, the reasoning for this conclusion is shaky. The minority states that it calculated this figure by first determining that 80 detainees were not subjected to EITs, and then finding that of this group interrogations of 34 detainees produced no intelligence reports. Unfortunately for the minority’s logic, there are plenty of reasons besides ineffective interrogations that could explain why the 34 detainees produced no intelligence. For example, the 34 detainees could simply have had no intelligence to provide. Here, the minority report fails to address that coercive tactics typically occur in an environment of incomplete data points. We don’t necessarily know if detainees have useful information. Even if we know this to a reasonable certainty, we still don’t know when they’ve given up all the information they have. EITs don’t necessarily improve the situation. As Stanford psychologist Philip Zimbardo showed long ago, once we put torture on the table, we quickly become confident that it will do the trick. That overconfidence, however, is part of the problem of imperfect knowledge, not part of the solution. Conscientious critics of the SSCI Report must also acknowledge that because of imperfect knowledge, we can’t know whether more traditional tactics, such as building rapport against the backdrop of legal detention, would be as effective as EITs over the long haul. The US experience with Somali terrorist Ahmed Warsame suggests that traditional interrogation can yield a great deal of useful information. Moreover, as CIA Director John Brennan stated in remarks on Thursday, traditional interrogation can gain that data without the reputational harm that the US suffered because of the use of EITs. If we are agnostic about whether coercive tactics yield actionable information, we should look for a tie-breaker. The best tie-breaker, as John McCain has said, goes not to torture’s efficacy but to its effect on us. Senator Lindsey Graham’s piece on waterboarding highlights the corrosive effect that coercive interrogation tactics can have on good order and discipline for those applying the coercion. Here, the CIA report uses bland language to describe a disturbing yet reassuring truth. According to the CIA, doctors observing the waterboarding of Zubaydah commented on the “lack of available data upon which to draw conclusions about [waterboarding’s] safety” (p. 6). Translation: The CIA’s own doctors worried that waterboarding would cause severe, lasting harm to subjects, and so informed their superiors. This disclosure is disturbing because the OLC opinions on waterboarding cited doctors’ attendance at the sessions as a factor in concluding that the technique was legal. A lawyer taking a less myopic view might have suggested to a client that it was a bad idea to use a technique that required a doctor’s attendance. However, the CIA’s disclosure is reassuring because the doctors’ reaction tells us that Zimbardo may have been partly wrong: for some of us, at least, brutalization is not inevitable, and a healthy revulsion is not so readily extinguished. That should give us some comfort that “who we are” is more resilient than some may have thought. The agnosticism about efficacy drawn from this debate demonstrates that ruling out EITs is a normative judgment, not a descriptive one. If we embrace that normative judgment while understanding the risks, our judgment will be more enduring. If we adopt that judgment glibly, while ignoring the mixed evidence on the efficacy of coercion, our judgment will not stand up to the “next attack.”