This is nasty work, I tell you, but it goes with the job. We didn’t try to come up with some sort of new idea for the guy. Everything happened like those clumsy films prepared him for, everything happened just like he expected: and this is always the surprise.
—IMRE KERTÉSZ, Detektívtörténet
On December 9, 2014, the Senate Select Committee on Intelligence released the long-anticipated, redacted summary of its report on the CIA’s detention and interrogation program. The 525-page document (hereafter report), with over 2,700 footnotes referencing secret CIA cables, emails, reports, interviews, and other documents, presents the Findings and Conclusions as well as an Executive Summary of the 6,700-plus-page full study (which remains classified) (United States Senate 2014a).1 Republican members of the committee, then in the minority, submitted their own minority report disagreeing with the majority’s findings (United States Senate 2014b). The CIA also responded with a redacted version of its June 2013 response to the report (Central Intelligence Agency 2014).
The game theoretic model on which this book is based, including the predictions that torture is ineffective yet more brutal than even proponents expect, was published in an academic journal in March 2012, more than two years before the release of the Senate report (Schiemann 2012a). The chapters you just finished reading were also largely written before the report’s release.2
The claim I made at the outset about the lack of systematic data still holds. The minority report by Republicans on the committee and the CIA’s objections in its own response demonstrate that there is no agreement even on what constitutes good evidence, so we still do not have the sort of rigorous test that can convince either side in the debate. Moreover, the sheer sizes of the summary and the responses—not to mention their complexities—prevent a full analysis here.
Nevertheless, even a brief examination is worthwhile. Bush administration and CIA officials have long claimed that the program was a success in terms of both information and adhering to the guidelines. The RIT model contradicts these claims. Given that the RIT outcomes and propositions reflecting the logic of interrogational torture were derived more than two years before the report was released, the report and the responses provide new and independent data to compare against both the Bush benchmarks and the RIT model’s predictions. Table PS.1 distills the differing predictions from previous chapters and pairs them side by side. They guide our examination of the report and the responses to it.
Table PS.1 PREDICTIONS FOR SENATE REPORT
|
Prediction |
Bush Model |
RIT Model |
|
Innocent detainees tortured |
No |
Yes |
|
Cooperative detainees tortured |
No |
Yes |
|
Frequency of torture |
Low |
High |
|
Severity of torture |
Minimum necessary |
Exceed limits |
|
Information |
Predictably reliable |
Unreliable, poor |
TORTURE OF INNOCENTS
The universe, or total population, of detainees in the CIA program was at least 119 (United States Senate 2014a, p. 40).3 Of these 119, nearly one in four (26 or 22%) were wrongfully held according to the CIA’s own guidelines and the President’s September 17, 2001, Memorandum of Notification authorizing covert action by the CIA (United States Senate 2014a, pp. 41–43, 109, 135–136, 154, 159, 458). These included:
1. Two innocent men detained as leverage against family members; one of these men was described by the CIA as “intellectually challenged”
2. Two innocent men falsely accused by KSM under torture
3. Two of the CIA’s own sources who had actually sought out the CIA to volunteer information
4. An innocent man falsely accused because of a “blood feud”
5. An innocent man the CIA later said was “at the wrong place at the wrong time”
6. Several cases of mistaken identity similar to that of Khaled El-Masri, sometimes based on translation errors
Two of these innocent men (Gul Rahman and Laid Ben Dohman Saidi, aka Abu Hudhaifa) were subjected to the enhanced torture methods. The report notes that this figure of 26 is also a conservative estimate insofar as it does not include those about whom there was internal CIA debate nor those who were later released when the CIA no longer considered them a threat. Nor does it include multiple detainees the CIA had detained by other countries because they did not meet the CIA’s criteria (United States Senate 2014a, pp. 41–43, 87).
The Republican members of the committee contest none of the foregoing. Indeed, their response is completely silent on the question of innocent men tortured by the CIA. Neither “innocent” nor any of its cognates and synonyms appear anywhere in the Republican report. Of the 26 wrongfully detained mentioned by the majority, the Republican response mentions only two of them by name, Haji Ghalgi (detained for leverage against a relative) and Hayatullah Haqqani (wrong place, wrong time) (United States Senate 2014b, p. 57). The Republican report says of the former only that he fabricated information and of the latter only that he did not provide fabricated information, both under “noncoercive interrogations.” It is unclear what the Republican committee members mean by “noncoercive,” since the rendition process and conditioning techniques—not to speak of the standard techniques—qualify as torture. The Republican report does include innocents implicitly—in the numbers of those who failed to generate information without the enhanced techniques. In other words, an innocent man’s failure to provide the information he does not have counts against the effectiveness of “noncoercive” interrogations!
For its part, the CIA admits in its official response that at least six detainees did not meet its own standard and were wrongfully detained (Central Intelligence Agency 2014, p. 75). Some of the other detentions, however, it claims, were “reasonable” under the standard at the time. This includes the “mentally challenged” brother who, after all, was released “in a matter of weeks,” even if the release of “[m]ost took three to six months” (Central Intelligence Agency 2014, pp. 52–53)).
The excuse offered by the CIA was that “the national priority was preventing attacks” (Central Intelligence Agency 2014, pp. 76–77)). Those of us old enough to remember those early months after 9/11 know that this is both absolutely true and utterly beside the point. To say that pressure to prevent an attack caused the detention (and torture) of innocents is to explain why it happened; it does not say that it did not happen.
In short, there is no doubt or argument that the CIA tortured innocent people, matching the RIT prediction in Proposition 12.1, not the Bush prediction. This violates the Innocents condition of the Torture Frequency benchmark. Evidence from the CIA program contradicts proponents’ standards and claims that innocents should not and will not be tortured.
TORTURE OF COOPERATIVE DETAINEES
This criterion would seem difficult to assess because, as we have repeatedly argued, one can never know whether a detainee has more information in his head. If he continues to hide information, then that torture is justified on the pragmatic model. If, however, the CIA itself comes to the conclusion that a detainee has no more information despite being tortured, then that torture is unnecessary according to the pragmatic model (and the CIA’s own standards). Ideally we would look at the case histories of the other 93 detainees who the Senate study agrees met the standard for detention and so actually had information to provide.
We do not, however, have this information systematically set out for any of them. Instead we have some information on some of them. If this information supports the benchmark—no Cooperative Detainees were tortured “unnecessarily” or “unjustifiably”—then it could still be the case that other detainee histories might not support the benchmark. But we would never know. If, however, there are significant instances of such unnecessary torture, then that is sufficient to indicate that the benchmark is not met. Consider two cases, Abu Zubaydah (United States Senate 2014a, pp. 51–63, 66–73, 230–236) and Abd al-Rahim al-Nashiri (United States Senate 2014a, pp. 92–99)).
Abu Zubaydah provided information “on al-Qa’ida activities, plans, capabilities, and relationships, in addition to information on its leadership structure, including personalities, decision-making processes, training, and tactics. Abu Zubaydah provided the same type of information prior to, during, and after the use of the CIA’s enhanced interrogation techniques” (United States Senate 2014a, p. 232). Zubaydah’s failure to provide “detailed and verifiable information on terrorist operations planned against the United States, including the names, phone numbers, email addresses, weapon caches, and safe houses” resulted in, after a 47-day isolation period, the “non-stop use of the CIA’s enhanced interrogation techniques” (combinations of shackling, hooding, diapering, nudity, slaps and attention grabs, walling, stress positions, waterboard, small and large confinement boxes [the latter presented as a coffin], white noise, dietary and temperature manipulation, sleep deprivation, and death threats) “24 hours a day” for nearly three weeks.
“After the use of the CIA’s enhanced interrogation techniques ended, CIA personnel at the detention site concluded that Abu Zubaydah had been truthful and that he did not possess any new terrorist threat information” (United States Senate 2014a, p. 71). “At no point during or after the use of the CIA’s enhanced interrogation techniques did Abu Zubaydah provide information on al-Qa’ida cells in the United States or operational plans for terrorist attacks against the United States” (United States Senate 2014a, p. 234). As a result of the failure to provide this information, the interrogation team flipped the rationale for using the techniques: “the interrogation team later deemed the use of the CIA’s enhanced interrogation techniques a success, not because it resulted in critical threat information, but because it provided further evidence that Abu Zubaydah had not been withholding the aforementioned information from the interrogators.” Neither the Republican nor the CIA responses contest the report’s claim that Abu Zubaydah never provided information under torture on cells or operational plans in the United States. (They do claim other information was provided, which we address below.)
Abd al-Rahim al-Nashiri provided information to a foreign government after his capture in the United Arab Emirates in mid-October 2002 including “details on multiple terrorist plots in which he was involved prior to his detention, [such as] the attacks against the USS Cole and the MV Limburg, plans to sink oil tankers in the Strait of Hormuz, plans to attack warships docked at ports in Dubai and Jeddah, and his casing of a Dubai amusement park” (United States Senate 2014a, p. 99). Intelligence reports with this information were disseminated throughout the CIA. Nevertheless, the next month he was rendered by the CIA first to the Salt Pit in Afghanistan for a short period before moving on to the black site in Thailand, where he was subjected to EITs, including waterboarding at least three times. When the Thai site was shut down, he was transferred with Abu Zubaydah to the CIA’s black site in Poland. There he was subjected to EITs during four periods: December 5–8, 2002, December 27, 2002, to January 1, 2003, January 9–10, 2003, and January 15–27, 2003.
The EITs were administered over the objections of the officers on the ground in Poland, who determined that al-Nashiri was cooperative and not resisting answering their questions (United States Senate 2014a, p. 498). Headquarters, however, was adamant: “[I]t is inconceivable to us that al-Nashiri cannot provide us concrete leads … When we are able to capture other terrorists based on his leads and to thwart future plots based on his reporting, we will have much more confidence that he is, indeed, genuinely cooperative on some level” (United States Senate 2014a, p. 94). The on-again, off-again pattern of EITs reflects this back and forth, with the local officers deeming al-Nashiri compliant after each period of EIT use and the headquarters criticizing them for making “sweeping statements” about al-Nashiri’s compliance and ordering more torture.
The torture included a range of authorized techniques as well as some unauthorized ones, including:
vertical shackling for two and one-half days, far exceeding the four hour maximum, racking a pistol near his head and running a cordless drill near him while blindfolded, telling him “[w]e could get your mother in here,” and “[w]e can bring your family in here,” pinning him down on his stomach with his head below his torso and forcibly injecting Ensure into his rectum, slapping him on the back of the head, keeping him nude during periods of vertical shackling, and putting him in improvised stress positions that required the intervention of a medical officer who feared al-Nashiri’s shoulders might start to dislocate (United States Senate 2014a, pp. 95–99, 513–514).
Al-Nashiri never provided the information that the CIA’s lead analysts at headquarters believed he possessed, namely “perishable threat information to help [the CIA] thwart future attacks and capture additional operatives” (United States Senate 2014a, p. 99). Contractor psychologist and interrogator Jessen himself concluded in October 2004, 21 months after the final documented use of torture, that al-Nishiri provided “essentially no actionable information” and “the probability that he has much more to contribute is low” (United States Senate 2014a, p. 99).
The CIA response does not challenge the claim in the report that al-Nishiri never provided threat information during or following the use of EITs. Al-Nishiri’s name does not even appear in the Republican response.
The report contains other instances of such unnecessary torture of cooperative detainees, but what we have found so far is sufficient to show that the CIA’s program tortured cooperative detainees long after they had no more information to provide. This confirms Implication 8.2 that cooperative detainees will be tortured for information they do not have and so matches the RIT model’s prediction, not the Bush model’s. In other words, the CIA program violated the Cooperatives condition of the Torture Frequency benchmark. Note that the failure to satisfy these conditions, along with those referring to innocent detainees above, entails a failure to satisfy the prediction of the ideal outcome in the Torture Justification Outcome benchmark. In other words, without having yet examined whether or not the program generated valuable information, we have already shown that the CIA program failed to live up to the ideal espoused by torture proponents.
TOTAL FREQUENCY OF TORTURE
The Total Frequency condition of the Torture Frequency benchmark predicts that the total frequency of torture will be low. We can assess this overall frequency of torture in the CIA program in several ways. Torture exceeds the threshold of low frequency if:
1. Any techniques were employed on those who did not “qualify” for the program (completely innocent or “wrongfully held” according to the CIA’s own standards) (Wrongful).
2. Unauthorized techniques were employed on those who did qualify (Unauthorized).
3. Authorized techniques were employed on those who did qualify without giving them a chance to cooperate or even after they had cooperated (Immediate).
4. Authorized techniques were employed on those who did qualify without the required approval for specific techniques (Unapproved).
5. The total number and proportion of detainees in program subjected to torture is high (Total Frequency).
We might also include longer or more frequent or intense use of authorized techniques against those who did qualify, but these are better discussed below in terms of torture severity.
First, we have already seen that nearly a quarter of detainees were subjected to some form of torture without cause because they were wrongfully held. Using torture on those who did not qualify according to the CIA’s own standards by definition means that it was used more frequently than permitted, satisfying condition Wrongful.
Second, the Senate study documents, and the CIA does not deny, that the agency used some unapproved techniques on detainees who did “qualify” for the CIA program, satisfying condition Unauthorized.
The torture of al-Nishiri described above is a clear example. The CIA apparently put “pressure on [an] artery” and performed “mock executions” on other detainees (United States Senate 2014a, p. 216). Abu Zubaydah was warned somewhat obliquely to “keep in mind” the “welfare” of his family, while contract torture psychologists Jessen and Mitchell threatened KSM much more directly. On KSM’s first day at the black site in Poland, they told him that “if anything happens in the United States, ‘[they] were going to kill [his] children’.” A picture of his sons hung up in his cell amplified their message (United States Senate 2014a, p. 513, also pp. 11, 111, 117).
The CIA never listed “rectal hydration” or “rectal feeding” in their list of techniques, and the CIA leadership continues to maintain with a straight face that they are “well acknowledged medical technique[s] to address pressing health issues” (Central Intelligence Agency 2014, p. 79). (The Republican minority response ignores rectal feeding/hydration entirely.) Rectal feeding was indeed “well acknowledged” as a last-ditch medical procedure—a century ago (Short and Bywaters 1913). Its medical use, however, faded into history along with mercury cures, drilling holes in the skull, bloodletting, and the urine treatment (don’t ask). As anyone who has visited a hospital or even seen one on TV knows, modern medicine uses either an IV infusion or nasogastric tubes (i.e., through the nose).
The reason “[r]ectal hydration is almost never practiced in medicine [is] because oral and intravenous routes of fluid administration are more effective … The large colon has the capacity to absorb fluids, but has a very limited capacity to absorb nutrients with the exception of glucose and electrolytes. Pureed food and nutritional supplements, such as Ensure, should never be administered rectally.”4 CIA medical officials, of course, knew this very well, with one medical officer at a black site considering normal “IV infusion as safe and effective.” The CIA’s chief interrogator stated frankly the real purpose: to demonstrate the interrogator’s “total control over the detainee” (United States Senate 2014a, pp. 126, 108).
Third, although the guidelines required and the CIA told Congress that the techniques would be applied only after detainees refused an initial opportunity to cooperate and only in a graduated fashion until they cooperated, the report found that “detainees who were subjected to the CIA’s enhanced interrogation techniques were usually subjected to the techniques immediately after being rendered to CIA custody” (United States Senate 2014a, p. 9). Indeed, by my count working through the report, a minimum of over one in four (10) were subjected to enhanced torture immediately, without being given an opportunity to cooperate. Some of these had even already declared to the CIA their intent to cooperate. Others had already provided valuable information prior to rendition (United States Senate 2014a, pp. 9–10, 103, 459), including nearly one in three subjected to EITs (12 of 39), satisfying condition Immediate.
In one case, the CIA had doubts about whether an Arsala Khan was the person they were seeking. The solution was to go straight to torture “to make a better assessment regarding [his] willingness to start talking, or assess if [Khan] is in fact the man we are looking for” (United States Senate 2014a, p. 458) (my emphasis). After a month of “extensive use of” EITs so that he could barely speak and had hallucinations his family members were being mauled by dogs, “‘the CIA concluded that […] Khan does not appear to be the subject involved in … current plans or activities against U.S. personnel or facilities,’ and recommended that he be released to his village with a cash payment.” He was nevertheless held by the U.S. military for another four years. The reason for his kidnapping and torture? He had been falsely accused by someone with a vendetta against him (United States Senate 2014a, pp. 42, 136).
Fourth, the CIA employed generally approved techniques on many detainees without the required authorization from CIA headquarters for that particular detainee. For example, the Senate study found that 17 of the 39 detainees who were subjected to the enhanced techniques had it done to them without the required headquarters approval—44% of all those receiving EITs (United States Senate 2014a, p. 127).5 Some of these techniques, such as nudity, dietary manipulation, and water dousing, were not approved at the time they were employed. In other cases, generally authorized techniques were used on detainees who had been ordered to undergo other enhanced techniques. In short, it is clear that the CIA program satisfied condition Unapproved.
Walling, for example, was employed against Abu Hazim in addition to other enhanced techniques such as dousing with water and sleep deprivation. Walling was not approved because Hazim had broken his foot during his capture. He was also forced into standing sleep deprivation, despite approval for seated sleep deprivation only. Abd al-Karim also broke a foot attempting to escape his captors and so was not supposed to have received some of the techniques. He was nevertheless subjected to cramped confinement, various stress positions, walling, and standing sleep deprivation (United States Senate 2014a, pp. 138–139)).
Fifth and finally, consider the last condition, Total Frequency. Begin by recalling that the total (minimum) population of detainees in the CIA program was 119. There were (again a minimum of) 39 detainees subjected to EITs (i.e., torture beyond the rendition/conditioning and the standard technique). Thus, on the most conservative (i.e., favorable to proponents) counting possible, one in three detainees in the program were tortured.
The problem is that the conditioning and standard techniques—isolation, nudity, sleep deprivation, various stress positions, cold water dousing, food manipulation (reduction), use of loud music or white noise, temperature manipulation, and the use of diapers—also constitute torture, as we saw in Chapter 2. Of course the rendition process itself, with sensory deprivation, nudity, and forcible sodomy, constituted torture in its own right. Assuming the standard rendition protocol was followed for all CIA captives, fully 100% of the detainees in the program were tortured in one way or another.
Any way you look at it, torture was very frequent, confirming the first Slippery Slope proposition (12.4) that once torture is admitted as an interrogation technique, it will exceed the limits and controls imposed on it and become more frequent than proponents expect. Thus, the results of the CIA program match the RIT model prediction and fail to match the Bush model prediction, violating the Torture Frequency benchmark. In terms of the frequency of torture, then, the CIA program lived up to neither the actual Bush program nor the ideal model of proponents.
TORTURE SEVERITY
Although the logic of interrogational torture requires more (brutal) torture to extract information, the pragmatic model’s Torture Severity benchmark argues that its severity will approximate the minimum degree necessary to compel valuable information. Torture proponents are a little vague on exactly what this means and, to be fair, as we said in Chapter 12, we cannot pin this down too closely. Even so, if there are clear-cut cases of torture beyond the program’s own limits, then this is a problem. This can happen in five ways:
1. Particularly brutal unauthorized techniques (Unauthorized)
2. Authorized techniques employed longer than permitted (Duration)
3. Authorized techniques employed more frequently than permitted (Frequency)
4. Authorized techniques employed more intensely or severely than permitted (Severity)
5. Authorized techniques employed in combinations violating guidelines (Combinations)
We have already seen the brutality of some of the unauthorized techniques, with rectal hydration/feeding at the top of the list. Five Detainees were subjected to rectal feeding/hydration. One of them, Mustafa al-Hawsawi, was later determined to have acquired “hemorrhoids, an anal fissure, and symptomatic rectal prolapse” (United States Senate 2014a, p. 126). Several detainees were told they would never leave the black sites alive (United States Senate 2014a, p. 11). Detainees were forcibly immersed in ice water baths or hosed with cold water while standing and shackled to the ceiling before being taken to cells with temperatures as low as 59 degrees Fahrenheit (United States Senate 2014a, p. 131). One CIA psychologist recalled hearing Abu Hudhaifa “gasp out loud several times as he was placed in the tub” of ice water (United States Senate 2014a, p. 445). Although the CIA reported to the Justice Department that lights were always on for security, in fact they often kept detainees in total darkness. This method they apparently discovered by accident. One day interrogators found Ramzi bin al-Shibh “cowering in the corner, shivering” after his cell light had gone out. Consequently, his interrogators “decided to use darkness as an interrogation technique.” He was then subjected to nude standing sleep deprivation, his hands shackled over his head, his feet manacled, “in total darkness” (United States Senate 2014a, p. 448).
In other cases, interrogators employed authorized techniques, but employed them in unauthorized ways: longer, more frequently, more intensely, or in combinations other than guidelines and interrogation plans permitted. The Senate report is replete with such instances. Consider just the following two cases:
Ridha al-Najjar was arrested in Karachi, Pakistan, in May 2002 and detained by a foreign government. Within weeks he offered up information on al-Kuwaiti, Bin Laden’s courier (United States Senate 2014a, p. 408). Nevertheless, he was rendered to CIA black sites, eventually ending up at the Salt Pit in Afghanistan. In early August 2002, the CIA began subjecting him to “loud music, nutritionally sufficient but poor food, sleep deprivation, and hooding” (United States Senate 2014a, p. 79). A little over a month later, his CIA interrogators deemed him “clearly a broken man … on the verge of a complete breakdown” and willing to do whatever the CIA wanted.
That, however, did not stop the torture. When a military legal officer visited in November, he noted that the interrogation plan for al-Najjar included “isolation in total darkness; lowering the quality of his food; keeping him at an uncomfortable temperature (cold); [playing music] 24 hours a day; and keeping him shackled and hooded.” He had also had his hands handcuffed over his head for 22 hours a day for two straight days, with a diaper substituting for a toilet (United States Senate 2014a, p. 79). The CIA produced one intelligence report from him following his detention and interrogations (United States Senate 2014a, p. 80).
Abu Ja’far al-Iraqi was transferred to CIA custody from the U.S. military in September 2005 (United States Senate 2014a, p. 174). On December 1, 2005, CIA Director Porter Goss authorized using EITs on al-Iraqi because the “CIA believes that Abu Ja’far possesses considerable operational information about Abu Mu’sab al-Zarqawi (the leader of Al Qaeda in Iraq).”6 As a result, al-Iraqi
was subjected to nudity, dietary manipulation, insult slaps, abdominal slaps, attention grasps, facial holds, walling, stress positions, and water dousing with 44 degree Fahrenheit water for 18 minutes. He was shackled in the standing position for 54 hours as part of sleep deprivation, and experienced swelling in his lower legs requiring blood thinner and spiral ace bandages. He was moved to a sitting position, and his sleep deprivation was extended to 78 hours. After the swelling subsided, he was provided with more blood thinner and was returned to the standing position. The sleep deprivation was extended to 102 hours. After four hours of sleep, Abu Ja’far al-Iraqi was subjected to an additional 52 hours of sleep deprivation, after which CIA Headquarters informed interrogators that eight hours was the minimum rest period between sleep deprivation sessions exceeding 48 hours. In addition to the swelling, Abu Ja’far al-Iraqi also experienced an edema [swelling] on his head due to walling, abrasions on his neck, and blisters on his ankles from shackles (United States Senate 2014a, p. 175).
It is easy to find more examples. KSM ingested so much water during waterboarding sessions that his stomach was “distended” and water gushed out of his mouth when interrogators pressed down on his stomach.7 At one point, a medical officer instructed interrogators to add salt to the water to prevent “water intoxication” and “electrolyte dilution.” A medical officer on site described the sessions as “a series of near drownings” (United States Senate 2014a, p. 112). At the Salt Pit, Gul Rahman was subjected to “48 hours of sleep deprivation, auditory overload, total darkness, isolation, a cold shower, and rough treatment [i.e., dragged naked and hooded through the hallways while being punched and slapped]” before being shackled to the wall in a way that forced him to lie on the bare concrete floor, dressed in nothing but a sweatshirt. He was found dead from hypothermia the next day (United States Senate 2014a, p. 80). The study found that “detainees who were subjected to the CIA’s enhanced interrogation techniques and extended isolation exhibited psychological and behavioral issues, including hallucinations, paranoia, insomnia, and attempts at self-harm and self-mutilation” (United States Senate 2014a, p. 11).
Consistent with its breezy indifference to the torture of innocents, the Republican response is also silent on the CIA program’s abuses in terms of severity. It makes no attempt to counter the claims made in the Senate report. There is no mention in the Republican response of hallucinations, water dousing, rectal hydration/feeding, threats to detainees and their families, or other abuses documented in the Senate report.
The CIA admits to having “erred” in the use of some techniques, plays down other violations as problems with the program early on which were corrected later, or charges the Senate report with leaving out “clarifying detail.” But the “most important” rebuttal is that Congress had been briefed about all this anyway (Central Intelligence Agency 2014, p. 78). The CIA response challenges only four claims about brutality: hallucinations from sleep deprivation, water dousing, rectal feeding/hydration, and waterboarding.
Hallucinations, the CIA argues, were rare and did not really matter anyway since they went away when detainees were eventually permitted to sleep. The agency’s response also simply repeated the claim that interrogators stopped sleep deprivation when detainees hallucinated, ignoring the Senate’s documentation of Hassan Ghul’s continued sleep deprivation despite hallucinations (United States Senate 2014a, p. 158).
Water dousing, the CIA admits, was a problem at the Salt Pit due to cold temperatures, but this was corrected later at other sites (Central Intelligence Agency 2014, p. 79). The CIA response does not address the documented cases where water dousing was employed in such a way that—in the words of one CIA interrogator—“can easily approximate waterboarding.” Indeed, Abu Hazim’s water dousing included a cloth over his mouth and water poured directly on his face. A medical officer removed the cloth when Hazim’s face “turned blue” (United States Senate 2014a, pp. 132–133)).
The CIA ignores rectal feeding entirely in its response. As we have seen, it claims that rectal hydration was both necessary and an established medical practice. A two-minute Internet search or brief conversation with any doctor will tell you that the claim that rectal hydration is “more efficient than a naso-gastric tube” is absurdly false. The claim that it was necessary for hydrating “noncompliant” detainees because needles and nasogastric tubes were dangerous is hard to square with the “total control” of the Detainees in theory and in practice (United States Senate 2014a, p. 513). It’s also hard to square with logic: Would you be more compliant with rectal hydration as compared to an IV or nasogastric tube? In any case, the Senate report points out that
the assertion [by the CIA in its response] that Majid Khan was “uncooperative” prior to rectal rehydration and rectal feeding is inaccurate. As described in CIA records, prior to being subjected to rectal rehydration and rectal feeding, Majid Khan cooperated with the nasogastric feedings and was permitted to infuse the fluids and nutrients himself (United States Senate 2014a, p. 141).
Khan would later engage in self-destructive behavior, including attempts to cut “his wrist,” “a vein in the top of his foot,” and the skin on his elbow “using a filed toothbrush.” He also tried to “chew into his arm at the inner elbow” (United States Senate 2014a, p. 141).
Finally, the CIA also admits that it used waterboarding in ways that deviated from the guidelines (but not the “principles” authorizing the technique). It mentions KSM’s waterboarding in particular and suggests that part of the problem was how “adept” he was at “resisting the technique” (Central Intelligence Agency 2014, p. 79).8 Indeed, one CIA cable from KSM’s detention at the Polish black site in March 2003 accused KSM of lying in order to get himself tortured!
[T] he enhanced measures resulting from his lying in [sic] details could be a resistance strategy to keep the interrogation from threatening issues … [KSM’s] apparent willingness to provoke and incur the use of enhanced measures may represent a calculated strategy to either: (A) redirect the course of the interrogation; or (B) to attempt to cultivate some doubt that he had knowledge of any current or future operations against the US (United States Senate 2014a, p. 118).
In any event, the CIA dismisses the relevance of the waterboard problems since “only three detainees” were waterboarded (Central Intelligence Agency 2014, p. 79). Unfortunately, there is strong evidence that, in fact, the CIA waterboarded additional detainees. First, there are the cases of water dousing becoming “indistinguishable” from waterboarding discussed above. Second, the CIA was “unable to explain” “a CIA photograph of a wooden waterboard … surrounded by buckets, … a bottle of unknown pink solution (filled two thirds of the way to the top), and a watering can resting on the [waterboard’s] wooden beams” at a time and detention site where no waterboarding was ever approved (United States Senate 2014a, p. 132). Finally, the Senate report gives credence to a September 2012 report by Human Rights Watch alleging the waterboarding of CIA detainee Mohammed Shoroeiya, aka Abd al-Karim, at the same detention site as the photograph (United States Senate 2014a, p. 133).9
In short, the CIA rebuttal fails to counter the obvious conclusion: The CIA program also slid down the second Slippery Slope (Proposition 12.5): Once torture is admitted as an interrogation technique, the strategic incentives facing the interrogator result in increasingly brutal forms of torture. This again matches the RIT model’s prediction and fails to meet the pragmatic model’s Torture Severity benchmark and thus the Severity condition of the Torture Justification Outcome as well. The CIA program, like those that went before it, could not escape the inevitable consequences of introducing torture into an interrogation room.
INFORMATION RELIABILITY
The last benchmark is the most important for torture proponents. Some of these proponents may be willing to accept the torture of innocents, of cooperative detainees, of torture surpassing all limits on frequency and brutality as long as it can be counted on to produce the necessary valuable information. Indeed, the only possible argument that can be made for the practice is that it works, that it satisfies the Information Reliability benchmark:
Benchmark 1 Information Reliability: Most detainees have information and give up (nearly) all of it so that the ratio of clear and valuable information to all information will be high.
Only if this prediction is met can interrogational torture be defended even in these proponents’ eyes.
The trouble is—even leaving aside the fact that it is far from clear that “most” detainees had valuable information—that this is the most difficult benchmark to assess. It would take an entire book of its own to trace out the claims and counterclaims about who knew what and gave up what part of it when and under which circumstances. Figuring how important some bit of information was on its own as compared to its place in the larger context compounds the problem. Nor do the redactions in the report and responses make this any easier.
Luckily we have another option, one similar to the strategy we adopted in assessing the other benchmarks. We need to see if there is any evidence of an obvious failure to satisfy the benchmarks. If we find instead that there are few instances of detainees providing no information or fabricated information under torture, if we find that only torture could reliably (predictably) generate reliable (good, actionable, previously unknown threat) information from these detainees, then the Information Reliability benchmark is satisfied by the data we have available. If the opposite is true, if the torture frequently resulted in no information or false information and not valuable information which was otherwise unobtainable, then that is sufficient to demonstrate that the CIA program failed to meet the pragmatic model’s own criterion of success.
No Information
Muhammad Rahim, the last detainee in the program, was captured in Pakistan on June 25, 2007. The CIA thought he would likely know information about Al Qaeda, including the location of Osama bin Laden, and rendered him to a black site in Afghanistan. When he “declined” to offer up any information on threats to the United States or the locations of Al Qaeda leaders, a team of four CIA interrogators began torturing him. The methods included facial slaps, abdominal slaps, the facial hold, eight different periods of extended sleep deprivation while shackled in a standing position in a diaper and a pair of shorts, and a diet of water and liquid Ensure. He provided “historical information” but nothing about the whereabouts of Al Qaeda leaders in questioning during his interrogation. At one point he was left isolated in his cell with “minimal contact” for six weeks. He still didn’t talk.
His interrogators reported back to headquarters that Rahim’s behavior had “demonstrated that the physical corrective measures available to [interrogators] have become predictable and bearable.” A high-ranking officer in the Counterterrorism Center thought the harsh measures were a bad idea since it made a detainee who withstood them to “believe [they] had won.” The CIA never disseminated any intelligence reports based on information from Rahim (United States Senate 2014a, pp. 189–194)).
The Abu Ja’far al-Iraqi case discussed just above provides another example. Here is what was originally put into a draft of the President’s daily briefing (PDB) from the CIA following his torture: Abu Ja’far al-Iraqi provided “almost no information that could be used to locate former colleagues or disrupt attack plots.” This statement was later deleted at the urging of an al-Iraqi interrogator, who worried “[i]f we allow the Director to give this PDB, as it is written, to the President, I would imagine the President would say, ‘You asked me to risk my presidency on your interrogations, and now you give me this that implies the interrogations are not working. Why do we bother?’ We think the tone of the PDB should be tweaked” (United States Senate 2014a, p. 175).
Even the poster boys for the CIA program, Abu Zubaydah and KSM, withheld some vital information at the same time they provided other intelligence. Both “denied any significant connection between al-Kuwaiti [a suspected courier] and [bin Laden],” though the CIA later determined that they had information and so had withheld it (United States Senate 2014a, p. 420). Abu Faraj al-Libi, Ammar al-Baluchi, and Khalid bin Attash also almost certainly withheld information they possessed on the suspected bin Laden courier (United States Senate 2014a, pp. 404–426)).
KSM provided information on a plot to target various U.S. and other Western locations and interests in Karachi, Pakistan, but only after confronted with evidence that two Al Qaeda operatives had been captured (Khallad bin Attash and Ammar al-Baluchi) (United States Senate 2014a, pp. 265–272)). During the period he was subjected to EITs, KSM had been silent on these plots. A bitter CIA cable said that the Al Qaeda unit at headquarters was “disappointed” and this “long-standing omission [was] a serious concern, especially as this omission may well have cost American lives had Pakistani authorities not been diligent in following up on unrelated criminal leads” (United States Senate 2014a, p. 272). KSM was joined by Ramzi bin al-Shibh, Ammar al-Baluchi, and Khallad bin Attash in withholding information about another plot to attack Heathrow Airport and Canary Wharf in London (United States Senate 2014a, p. 326). And so on.
These cases illustrate a broader finding of the Senate investigation: Seven of the 39 detainees subjected to the enhanced torture methods provided no information whatsoever during their time in CIA detention. In other words, the CIA disseminated exactly “zero intelligence reports … based on information” these seven detainees provided under torture (United States Senate 2014a, pp. 9, 25). Considering only this figure, the enhanced torture techniques, the ones the United States really needed to use to get vital information otherwise unobtainable, were 82% reliable in getting any kind of information out of a detainee worth reporting (and we shall see that a good deal of it will turn out to be false). That’s a B−. Perhaps this is acceptable if you’re just looking to pass Psych 101 and graduate, but would it be acceptable if it were your car’s reliability? Would you fly an airline that crashes “only” 18% of the time? Whether or not the United States of America should torture people seems closer in importance to the airline example than it does to the “passing Psych 101” example.
The Republican response reflects a different view, opining that a 82% success rate “sounds pretty good” (United States Senate 2014b, p. 55). In any case, they say, it is better than the 57.5% success rate they calculate for the 80 detainees not subjected to EITs. Even if this constituted an accurate comparison (it does not, as we shall see), the relative comparison is a disturbing defense of torture. Is simply doing better good enough to justify torture? How much better is better enough?
More problematic is the fact that the 57.5% is an inaccurate figure for comparison, for three reasons. First, either bizarrely or disingenuously, this calculation reflects the failure of innocents subjected to other, non-EIT torture methods, to provide the information they never had! Of the 80 detainees not subjected to EITs, 34 did not provide information, generating the Republican 57.5% effectiveness rate. The problem is that this 80 includes at least 24 detainees who were wrongfully held (the other two wrongfully held detainees were subjected to EITs). This means that they did not have the information the CIA thought they did—making it impossible for them to have provided good information. (Of course they could have provided false information, but the same is true of those who supplied information under EITs. We return to information quality below.) These 26 innocent detainees should be dropped from the Republican calculation (24 from the no-EITs group and two from the EITs group).
More precisely, we should drop them from the count of the detainees in each group (no-EITs and EITs) who provided no information. Of the 39 detainees subjected to EITs, seven provided no information. Since two of these 39 were innocent, it stands to reason that we should drop them from the seven who did not provide information. This reduces the number of non-information-providing detainees in the EIT group to five (from seven).
We do the same thing for the group not subjected to EITs. According to the Republican response, 34 of the 80 in this group did not provide information. From the total of 26 innocents, 24 were in this non-EIT group (the other two were subjected to EITs). Again, it stands to reason that these 24 innocents could not have provided information and so should be subtracted from the non-information-providing group of 34, leaving 10. Table PS.2 sets out these corrected figures in parentheses, along with the corrected success and failure rates calculated using the method in the Republican response.
Table PS.2 CORRECTED INFORMATION RATES
As you can see, the effectiveness rate of EITs went up a bit, from 82% to 86%, as a result of dropping the two innocent detainees from the “no information” column. Once, however, the innocent detainees who by definition could not have provided information are subtracted from the “no information” count of the non-EIT group, the success rate of non-EITs shoots up to the old EIT proportion of 82%. These rates are both commonsensically and statistically indistinguishable.10
Second, more importantly, as we have already shown, the 80 detainees who did not receive EITs were still tortured, just not tortured to quite the same degree. The division into “enhanced,” “standard,” and “conditioning” is an arbitrary one. Indeed, some techniques were initially not on any list, then classified as standard, and then became enhanced. We might just as well break out the numbers by rectal feeding or waterboarding, since those were the worst of the worst in many eyes. The point is that everyone in the program was forcibly sodomized, hooded, manacled, kept nude, and had their diets manipulated to one degree or another. The appropriate comparison, then, would actually be to entirely noncoercive interrogations, such as those conducted by military and FBI interrogators throughout American history, including against Al Qaeda in Iraq and Afghanistan. Unfortunately we do not have these data (and anyway they would suffer from the same epistemological and empirical problems we covered in Chapter 2).
Third, this counting method also ignores the fact that the detainees receiving EITs were more likely to have more, and more valuable, information. That is why they qualified for enhanced torture after all. If detainees were equally likely to talk (i.e., whether or not they had been tortured with the enhanced techniques, as we saw in the corrected figures just now), then it is not surprising at all that those receiving EITs would generate more reports. By definition, they had more information to give. While it is no doubt true that they may have wanted to hide that information even more, it is just as true that your ability to hide information is unrelated to, independent of, how much information you happen to have.
Suppose we set all these problems aside and compare the total number of detainees who provided information resulting in disseminated reports to those associated with no reports. If we do this, the overall success rate of the program drops to 66% (78 of 119). That’s a solid D. Keep this up and you won’t graduate.
The Republican response rightly points out that the number of reports is less important than the quality of what was provided. In their words, the “true test of effectiveness is the value of what was obtained—not how much or how little was obtained” (United States Senate 2014b, p. 55). They make this point to dismiss the Senate’s claims about the failure rate (18% of EITs, 33% of the total). Fair enough. Before we turn to the quality of the information coming out of the CIA program, however, note that this shift in attention does not entirely do away with the problem of those providing no information.
What about the value of the information which was never elicited from those who withheld it? An argument justifying torture as the only possible way to get “vital … otherwise unavailable actionable intelligence” cannot simply dismiss those cases in which the techniques failed to get that intelligence (United States Senate 2014a, p. 322). It may be that the information kept hidden by detainees who don’t reveal information is much more valuable than the information which is obtained. Maybe not. The point is we do not know.
In any event, Republicans are absolutely right that it is important to look at information quality in addition to information quantity. Information can be good (unique, actionable intelligence on important threats or key enemies), neutral (nonuniquely, corroborative, confirmatory of out-of-date information), or bad (fabricated or false information which misleads or misdirects attention away from true plots, persons, locations, etc.). Obviously good is what interrogators are after. Neutral can be helpful, but is just as obviously of lower value. Moreover, proponents and the pragmatic model (rightly) do not justify torture on the basis that it merely provides corroborative or historical information. Bad is doubly bad. Bad information not only means not getting the vital, actionable threat information, but also entails actually moving further away from it, devoting time and resources in a direction which will ultimately prove fruitless and provide a greater advantage to the enemy. A reliable and effective interrogational torture program should have lots of good information, some neutral information, and very little bad information. Let’s start with the latter.
False Information
Based on a misreading of information from another tortured detainee, CIA headquarters directed the interrogators in Poland to press KSM on his plot to recruit African-American converts in the United States for terrorist attacks.
On March 21, 2003, KSM was waterboarded for failing to confirm interrogators’ suspicions that KSM sought to recruit individuals from among the African American Muslim community. KSM then stated that he had talked with Issa about contacting African American Muslim groups prior to September 11, 2001. The next day KSM was waterboarded for failing to provide more information on the recruitment of African American Muslims. One hour after the waterboarding session, KSM stated that he tasked Issa “to make contact with black U.S. citizen converts to Islam in Montana,” and that he instructed Issa to use his ties to Shaykh Abu Hamza al-Masri, a U.K.-based Imam, to facilitate his recruitment efforts (United States Senate 2014a, p. 294).
This (and other fabricated) information “required extensive FBI investigations.” Less than two months after providing this threat “information,” the FBI hosted a conference on the investigations following up on KSM’s leads, including attempts by Al Qaeda to recruit terrorists in the United States (United States Senate 2014a, p. 511). A little over a year later in June 2003, perhaps confronted with the FBI’s failure to turn up anything in Montana, KSM acknowledged that he fabricated the story about Abu Issa al-Britani and Montana because he was “under ‘enhanced measures’ when he made these claims and simply told his interrogators what he thought they wanted to hear.”
CIA headquarters, however, considered KSM’s retractions to be “another resistance/manipulation ploy” and “convenient excuses” and thus instructed CIA interrogators at the Polish black site “to get KSM to reveal … the key contact person in Montana.” By the end of the month, the bin Laden unit at CIA headquarters finally agreed that KSM’s reporting about African American Muslims in Montana was “an outright fabrication.” And, indeed, “no individuals related to KSM’s reporting were ever identified in Montana” (United States Senate 2014a, pp. 294–295)). The FBI had wasted precious resources on a wild goose chase as a result of KSM’s tortured false information.
The Senate report is full of other information fabricated by KSM, including some which led to the capture and detention of innocent men, and is worth examining (United States Senate 2014a, e.g., pp. 240–241, 296, 511). How about other detainees? Following two days of enhanced torture in May 2003, Ammar al-Baluchi told his interrogators that he had provided false information the previous day (United States Senate 2014a, p. 414). Hambali, a Southeast Asian operative captured in the summer of 2003, provided information and was deemed cooperative from the beginning but was nevertheless subjected to enhanced torture. Within four months he had recanted his earlier information. The CIA itself believed those retractions (United States Senate 2014a, pp. 134–135)). The reason?
Hambali reported that
through statements read to him and constant repetition of questions, he was made aware of what type of answers his questioners wanted. [Hambali] said he merely gave answers that were similar to what was being asked and what he inferred the interrogator or debriefer wanted, and when the pressure subsided or he was told that the information he gave was okay, [Hambali] knew that he had provided the answer that was being sought (United States Senate 2014a, pp. 134, 282–283).
In other words, the CIA had engaged in leading questioning and got the answers it wanted. The result was false information.
Abu Bakr al-Filistini, aka Samr al-Barq, was being subjected to enhanced torture on August 1, 2003, under suspicion of involvement in a plot to make anthrax. He nevertheless told his interrogators that “we never made anthrax.” His interrogators were not happy and told him “that the harsh treatment would not stop until he ‘told the truth.’” “[C]rying … al-Barq then said ‘I made the anthrax.’” When his interrogators asked if he was lying, he admitted he was. “After CIA interrogators ‘demonstrated the penalty for lying,’ al-Barq again stated ‘I made the anthrax’ and then immediately recanted, and then again stated that he made anthrax. Two days later, al-Barq stated that he had lied about the anthrax production ‘only because he thought that was what interrogators wanted”’ (United States Senate 2014a, p. 109).
With respect to the CIA’s main claim that enhanced torture provided the information which led to bin Laden, the Senate found that
information from CIA detainees subjected to the CIA’s enhanced interrogation techniques—to include CIA detainees who had clear links to Abu Ahmad al-Kuwaiti based on a large body of intelligence reporting—provided … intentionally misleading … fabricated, inconsistent, and generally unreliable information on Abu Ahmad al-Kuwaiti throughout their detention (United States Senate 2014a, pp. 405, 413–419, 425).
There were other cases where the information provided should have been suspect. The torture of Arsala Khan resulted in one intelligence report which was based on information he provided while hallucinating that his interrogators had given his family to dogs who were mauling and killing them (United States Senate 2014a, p. 135). Muhammed Rahim threatened to provide false information if his interrogators continued to torture him (United States Senate 2014a, pp. 190–191)).
The CIA response devotes surprisingly little space to the fabricated or otherwise inaccurate information generated from torture. In the only specific case it addresses directly, the CIA claims that Hambali told the truth under torture and it was his recantation which was a lie (Central Intelligence Agency 2014, p. 50). The Senate examined these claims and found that “the CIA’s June 2013 Response is incongruent with the assessment of CIA interrogators at the time,” quoting from the CIA’s own cables (United States Senate 2014a, pp. 283, 509). The CIA’s more general defense is that they put disclaimers on reports dissemination information from detainees and issued retractions when appropriate (Central Intelligence Agency 2014, pp. 51–55)). These caveats and disclaimers, however, apparently did not prevent the CIA, FBI, and other agencies from chasing down false leads nor prevent the CIA from arresting innocent people. The retractions do not mention that the fabricated information came from torture.
The CIA response does, however, point to a successful uncovering of fabricated information via torture. In July 2004 Janat Gul was rendered into CIA detention on the basis of a CIA source who fingered Gul as having information on Al Qaeda plans to launch a pre-election attack against the United States. The CIA subjected Gul to enhanced torture, “including continuous sleep deprivation, facial holds, attention grasps, facial slaps, stress positions, and walling.” He eventually “experienced auditory and visual hallucinations … [telling] CIA officers that he saw ‘his wife and children in the mirror and had heard their voices in the white noise.’” He later “asked to die, or just be killed.” Although his interrogators began to believe Gul did not possess the information, headquarters ordered the enhanced torture be continued. He never provided any threat information (United States Senate 2014a, pp. 161–163)).
But, the CIA claims, this is actually a success. Why?
Because Gul’s denials of knowledge under torture of any threat helped the CIA figure out that its original source had fabricated information. In other words, torturing someone without information is fine as long as it helps uncover the lies of the people who identified the tortured person.
The Republican response dismisses the importance of false and fabricated information more explicitly. “Fabrication is simply not a good measure of ‘effectiveness,’ because detainees are often strongly motivated to protect the identities of their terrorist colleagues and the details of their terrorist operations” (United States Senate 2014b, p. 57). For a response smugly scornful of the Senate report’s “logical” flaws and “faulty premises,” this is an odd locution (United States Senate 2014b, pp. 55–58)). It says that because detainees have strong motivation to protect information, false information is a bad measure of effectiveness. This makes no sense. False information’s validity as a measure of effectiveness has to do with what we mean by “effective” (actionable, critical, etc.); the motivations inside the heads of detainees have no bearing on that meaning.
What the Republican response is trying to say is the same thing the CIA said: We should not be surprised; detainees lie whether or not you torture them.11 This is, of course, true, but it undermines the rationale for the CIA torture program, which is that the methods were the one way to obtain “otherwise unavailable actionable intelligence” (for a list of the many occasions on which the CIA used this and cognate justifications, see United States Senate [2014a pp. 198–200, footnotes 1,050 and 1,051; pp. 321–322, footnote 1,667, and section III of the report]). The claim of torture proponents is that torture is necessary precisely because it gets us the information we would otherwise not get. Not the false and fabricated information we get from nontorturous methods, but the actionable intelligence we need. That is what “effective” means. The CIA also moves the goalposts much closer to the ball when it rewrites history to say that “[t]he purpose of the program was to minimize what was withheld with the understanding that obtaining complete disclosures from detainees in every case was not possible” (Central Intelligence Agency 2014, p. 51).12
The Republican response also attempts to defend the frequency of false information with the averral that “fabricated information can sometimes turn out to be highly significant” (United States Senate 2014b, p. 57). This is logic unencumbered by reason. First, the CIA in particular and interrogators more generally often do not figure out until (much) later that the statements were fabrications. In the meantime the information is taken at face value, with all that implies, which means that, at least for a while, fabricated information is a problem. Second, suppose we grant that “sometimes” (the right combination of) fabrications will (eventually) turn out to be significant. Do we start torturing knowing full well we will get bad information but hoping that “sometimes” the lies they tell us might somehow make sense later? Is this a sufficiently strong reason to torture? It is certainly far weaker than the justifications originally provided by the CIA and its proponents, let alone those of philosophers and legal scholars.
Good Information … Just not from Torture
Now it could be that despite the problems with no information and fabricated, false information, there was so much good information generated from torture in terms of both quantity and quality that it overwhelmed these problems. In order to assess this question, the Senate selected “20 of the most frequent and prominent examples of purported counterterrorism successes that the CIA has attributed to the use of its enhanced interrogation techniques.” The committee examined the contemporaneous cables, emails, intelligence reports, and other documents to assess whether or not it was the CIA’s use of enhanced torture which resulted in “unique, otherwise unavailable intelligence that led to the capture of specific terrorists and the ‘thwarting’ of specific plots” (United States Senate 2014a, pp. 249–251)). This review identified the following problems (United States Senate 2014a, pp. 9–10)):
1. In some cases there was no relationship between the cited counterterrorism success and any information provided by detainees during or after the use of the CIA’s enhanced interrogation techniques.
2. In the remaining cases, the CIA inaccurately claimed that specific, otherwise unavailable information was acquired from a CIA detainee as a result of the CIA’s enhanced interrogation techniques, when in fact the information was either
(a) corroborative of information already available to the CIA or other elements of the U.S. intelligence community from sources other than the CIA detainee, and was therefore not otherwise unavailable, or
(b) acquired from the CIA detainee prior to the use of the CIA’s enhanced interrogation techniques.
3. Examples provided by the CIA included numerous factual inaccuracies.
4. The CIA consistently omitted the significant amount of relevant intelligence obtained from sources other than CIA detainees who had been subjected to the CIA’s enhanced interrogation techniques— leaving the false impression the CIA was acquiring unique information from the use of the techniques.
5. Some of the plots that the CIA claimed to have “disrupted” as a result of the CIA’s enhanced interrogation techniques were assessed by intelligence and law enforcement officials as being infeasible or ideas that were never operationalized.
The Republican minority on the committee and the CIA, of course, dispute these claims. Partly they do so, as we have seen, by simply moving the goalposts and redefining what counts as good information, on occasion in Orwellian terms. Thus, corroborative information, repeating information previously provided while not under torture, and even false information are counted as good and useful. But only in part. The Republican and CIA responses also challenge the committee’s report on minute details of what Detainees revealed when and under what circumstances.
Sorting out who is right is extraordinarily difficult. The devil really is in the details of the combined 828 pages and 3,623 footnotes. Consequently, we examine here just one example illustrating the contested nature of claims about reliability: the capture of Jose Padilla.
Jose Padilla is a former Chicago gang member who converted to Islam in prison. He was later radicalized and in 2001 he traveled to Pakistan and Afghanistan to wage jihad and trained in explosives and other terrorist methods. On May 8, 2002, he was arrested upon landing in Chicago.
What was he plotting and how was he identified?
Initially he was thought to be plotting some sort of nuclear attack. The fact is, the attack plan was dismissed by the CIA itself before Padilla was even arrested as “cockamamie” (the plan called for “placing liquid uranium hexaflouride [sic] in a bucket, attaching it to a six foot rope, and swinging it around [the] head as fast as possible for 45 minutes,” something a physicist told the CIA would kill anyone trying to do it). Nevertheless, the CIA admits that it improperly continued to refer to this “Dirty Bomb” plot as late as 2007 in connection with Padilla (United States Senate 2014a, p. 252; Central Intelligence Agency 2014, p. 44). What was eventually discovered was that he was sent to the United States to try and blow up buildings with natural gas. The CIA has claimed that Padilla’s identity and involvement in this plot were discovered only after the CIA tortured Abu Zubaydah with the enhanced techniques (United States Senate 2014a, pp. 252–253)).
According to the committee’s report, “the CIA’s enhanced interrogation techniques played no role in the identification of Jose Padilla or the thwarting of the Dirty Bomb or Tall Buildings plotting.” In addition to the fact that the CIA already had information on Padilla sufficient to stop him and his plotting and the information Zubaydah supplied was not sufficient by itself to locate Padilla, the committee asserts that Abu Zubaydah “provided this information to FBI special agents who were using rapport-building techniques, in April 2002, more than three months prior” to the application of enhanced torture methods in August (United States Senate 2014a, pp. 256–257)).
In its June 2013 response, the CIA contested the committee’s claim that their existing information on Padilla was sufficient to have apprehended him without the additional information provided by Zubaydah (Central Intelligence Agency 2014, p. 44). Although, as the “chief of the Abu Zubaydah Task Force” herself put it in 2002, “AZ’s info alone would never have allowed us to find [ Jose Padilla],” it seems reasonable to conclude that the information was of some value. The real question is how it was procured, and the CIA also challenged the committee’s claim that Zubaydah had not been subjected to EITs when he provided the information. The agency noted that interrogators had already subjected Zubaydah to sleep deprivation (Central Intelligence Agency 2014, p. 88). The Republican response adds to this by asserting that Zubaydah’s treatment also “included nudity, liquid diet, [and] sensory deprivation” (United States Senate 2014b, pp. 68–69)). They cite sleep deprivation in particular as “play[ing] a significant role in Abu Zubaydah’s identification of Jose Padilla” (United States Senate 2014b, pp. 15–16)).
In other words, much hinges on exactly what was done to Abu Zubaydah leading up to when he provided the information on Padilla. Here is the apparent sequence of events, reconstructed from all three sources: United States Senate (2014a), pp. 54–56, 251–264, United States Senate (2014b), pp. 68–70), and Central Intelligence Agency (2014), pp. 87–88).
On the evening of April 15, 2002, the CIA team transferred a sedated Abu Zubaydah back to the black site from the hospital. He woke up at 11:00 pm. Constant questioning in rotating shifts of CIA and FBI agents kept him awake for the next 76 hours. On April 17 he provided information to his FBI interrogators about Al Qaeda, KSM (whom he had previously identified to FBI agents as the “mastermind” of the 9/11 attacks), and “general information on extremists in Pakistan.” At 3:00 am on April 19, 2002, he was permitted to sleep for three hours because he could no longer focus on the questions submitted to him and was providing “incoherent” answers. Interrogators woke him up at 6:00 am and began questioning him again.
Forty hours and thirty minutes later, at 10:30 pm on April 20, 2002, the FBI team began their interrogation shift. They allowed Zubaydah to sleep for two hours as well as eat, pray, and receive a medical checkup. Shortly thereafter, they confronted Zubaydah with emails showing that Zubaydah had sent two men to KSM. Zubaydah then told the FBI that two men had come to him with their plot to blow up a nuclear device in the United States. Zubaydah dismissed it as implausible (Padilla would later say that Zubaydah “chuckled” at the idea), but sent them to KSM for something that might work. Although he did not know their names, he provided physical descriptions of the pair. This information was similar to other reporting they had on Jose Padilla and within hours the CIA cabled back these connections to headquarters. The United States issued a travel alert for Padilla; and when he boarded a plane in Zurich two weeks later, an undercover FBI special agent boarded with him, watching him on the flight to Chicago, where they arrested him.
So what can we make of this? In terms of pure lack of sleep, two hours of sleep in a little under two days is not a lot, but it’s not too different from pulling a couple of all-nighters the night before a paper is due. Indeed, a CIA cable from the black site compared the “limited sleep” (note, not sleep deprivation) to that endured by medical students (United States Senate 2014a, p. 256). Moreover, Zubaydah had slept shortly before providing the information to the FBI and not to the CIA interrogators of the previous shift, before he slept. One might just as easily conclude that it was sleep, and not sleep deprivation, which generated the information.
In any case, the limited sleep regimen imposed on Abu Zubaydah in the second half of April 2002 was certainly a far cry from the degree of sleep deprivation that he would face four months later, in August. It was also very different from what is implied by the Republican response, which emphasizes the total amount of sleep and wakefulness over the entire April 19–21, 2002, period. This conceals the fact that Zubaydah gave up the information after having just slept.
It is also important to point out that the way in which Zubaydah was kept awake differed greatly from what would follow. He was kept awake by continuous questioning, not vertical or other shackling in stress positions hooded in a diaper, let alone cold water dousing or other techniques used to keep detainees awake. Nor was nudity or dietary manipulation the same as would later be the case. He was covered by a towel while the FBI questioned him. They gave him Coke and tea (United States Senate 2014a, p. 256). Finally, the questioning process itself was much different than the ways in which interrogators would later question detainees during the techniques, while shackled and perhaps hooded, dousing them with water, slapping them, walling them.
In short, the “enhanced” techniques to which Abu Zubaydah was subjected were very different from those eventually earning the euphemism. There is no denying they departed from normal FBI and military practice. There is also no denying, however, that they fail to qualify as an adequate example of the torture eventually practiced. If the Senate report was, as the Republican response argues, too “absolute” in its claims of ineffectiveness in this case, then the Republican response commits the same sin. The circumstances are far too ambiguous—at best—to qualify as the “direct refutation” claimed for it (United States Senate 2014b, p. 70). No “fatal flaw” or “analytical chain reaction” ensues (United States Senate 2014b, p. 62). Indeed, a senior CIA officer stated that “Padilla and the dirty bomb plot was prior to enhanced and he never really gave us actionable intel to get them” (United States Senate 2014a, p. 261).
Moreover, both the CIA and Republican responses, and even the Senate’s own report, fail to mention several other relevant facts explaining Zubaydah’s cooperation on the 21st despite his treatment. First, adopting the Republican’s more “chronological” “analytical methodology,” Zubaydah had already been cooperating with the FBI agents at the site and in the hospital before the CIA began torturing him upon his return from the hospital on the 15th (United States Senate 2014b, p. 61). He started to clam up under the new regimen. When he did offer information, it was to the FBI, using their questioning approach, not the CIA’s, as even the Republican response admits. Zubaydah recognized a photograph of Ramzi bin al-Shibh on the 18th and provided the name to his FBI interrogators (United States Senate 2014b, p. 72). After the FBI left, the CIA itself deemed Zubaydah less compliant. Indeed, they considered him so uncooperative they put him in isolation for a month and a half before subjecting him to full-blown enhanced torture including waterboarding for 19 straight days.
As a matter of fact, if the techniques were so necessary and effective, this is where one would expect to find the CIA and the Republican response listing off the new unique and valuable information gained. What do we find?
Although the Republican response claims that the August enhanced torture yielded “a significant amount of important information” from Zubaydah, they provide only two examples, one from August 20 and one from August 21, a few days before the CIA stopped using the enhanced torture methods.
On August 20, Zubaydah was asked how he would get back in touch with senior Al Qaeda operatives if he were released. He replied that he would seek out a “well-known associate of Hassan Ghul.” This associate could put him in touch with Ghul and other senior Al Qaeda associates (United States Senate 2014b, p. 74). The CIA claims that this was “unique information” without which the capture of terrorist Ramzi bin al-Shibh “would not have occurred” (Central Intelligence Agency 2014, pp. 106–107)). The Republican response echoes this interpretation, arguing that there is a “direct causal connection” between this information and the accidental capture of bin al-Shibh in a raid by Pakistani authorities (United States Senate 2014b, p. 75).
“Direct causal connection” is too strong by half. First, for the causal distance from the information to the ultimate capture of bin al-Shibh, the number and length of the causal links matter. There is a direct causal link between the Big Bang and my sitting at my desk writing these words, but it’s pretty distant. Second, and more importantly, the connection is not direct at all. As the Senate report documents in detail, both the CIA and the Pakistani authorities were already well aware of the associate’s connection to Ghul and had even already raided his house and questioned him. Zubaydah’s corroboratory—not at all “unique”—information simply confirmed the associate’s importance. Moreover, the Pakistani safe house raids did not target bin al-Shibh; he was picked up fortuitously. That’s not exactly “direct” (United States Senate 2014a, pp. 347–350)).
On August 21, 2002, CIA interrogators showed Zubaydah several photos of suspected terrorists. He recognized Ramzi bin al-Shibh and said that he worked with KSM. The CIA labeled this “significant new details” in cables at the time (United States Senate 2014a, p. 76). The problem is that there wasn’t much new to the CIA in this information. Not only was bin al-Shibh’s close association with KSM already known to the CIA from other sources, but Zubaydah himself had already told his interrogators about their association (United States Senate 2014a, p. 346). Even the Republican response’s own narrative demonstrates this. They point out that on May 19 and 20, before the 47 days of isolation and 19 days of brutal torture in August, Zubaydah had (1) identified another photo of him as “al-Shiba” and (2) stated that he was always with KSM. In the words of the Republican response, Zubaydah simply “confirmed” under torture what he had already said without the enhanced torture back in May (United States Senate 2014b, pp. 72–73)). There was no previously unknown, new information here, let alone anything actionable. As “[p]articipants in the interrogation of Abu Zubaydah” put it, Abu Zubaydah “probably reached the point of cooperation even prior to the August institution of ‘enhanced’ measures—a development missed because of the narrow focus of the questioning. In any event there was no evidence that the waterboard produced time-perishable information which otherwise would have been unobtainable” (United States Senate 2014a, p. 234).
In short, we have seen the failure of detainees to provide any information at all, to provide false and misleading information, and to provide some corroborative information, including information that confirms what the same detainee had already provided without torture. The examples cited here are not the only instances. The Senate report is full of others. And while the CIA and Republican responses are able to point to some minor and corroborative information provided under torture, there is no instance anything near to the sort of information supporting the pragmatic model’s claims of reliably generating reliable (valuable) information on imminent threats. Thus, the Bush program failed to meet the Information Reliability benchmark—the only possible justification for interrogational torture.