In his book, The Black Banners, former FBI Agent Ali Soufan describes multiple occasions when FBI and CIA reporting on a subject did not match. For example, he describes how his reporting and that of a CIA officer, Fred, differed during the investigation into the Millenium Plot.
My problems with him started within the first couple of days, after Pat D’Amuro received a phone call from FBI headquarters saying that my reporting of intelligence and Fred’s reporting of the same event didn’t match up.
An investigation was done and the Jordanians were consulted, and all concerned were advised that my reporting was correct and Fred’s was faulty.
Because of his flawed analysis, a total of twelve [redacted]–intelligence reports–had to be withdrawn. If portions of a cable are shown to be inaccurate, the entire cable is viewed as unreliable and suspect. (138-39)
Soufan elsewhere claims “there were discrepancies between information that went through CIA channels and what was reported in FBI channels” in some other cases. (119)
Adnan Latif’s redacted petition for cert has been released. The petition–plus the exhibits submitted with it–show that similar problems plagued at least one pair of reports on Latif. And those discrepancies, by themselves, prove that giving government interrogation summaries the presumption of regularity is untenable.
The pair of reports are DOD and FBI summaries of an interrogation of Latif conducted on May 29, 2002 (see PDFs 91 and 93-94) . As I noted in this post, even Latif’s factual return made it clear there were discrepancies between the two reports (though the unredacted parts of the factual return didn’t admit they recorded the same interview). The petition summarizes these discrepancies.
The reports, however, have numerous discrepancies. For example, one states that he is a Yemeni, App. 223a, while the other says both that he “claims Bangladeshi citizenship” and is a member of a Yemeni tribe, App. 221a. One says that he attended secondary school for “two or three years, and eventually graduated,” App. 223a, while the other states that he claimed to have “never graduated from high school,” App. 221a. It is obvious that at least one or perhaps both documents failed accurately to report what the translator was telling the interrogators.
The Bangladesh claim, incidentally, appears to derive from just one report, Latif’s Knowledgability Brief from February 2002 (which was not cited in his Gitmo file); his intake form (PDF 33-34) from December 31, 2001 clearly identifies him as an Arabic speaking Yemeni and notes he claimed he was picked up because he was an Arab. So it appears (though we can’t be sure) the DOD report writer wrote what would be consistent with the KB (and cited it), while the FBI report recorded what Latif said in the interview.
One more important discrepancy between the DOD and FBI reports from May 29, 2002: the DOD report says Latif was 16 when he suffered his head injury. The FBI report said he was 14. Latif’s factual return cites the differing ages as proof he kept changing his story (something similar happened in one of his CSRTs, but the confusion arose from his sense of time); but clearly here it was a difference of reporting, not of his report.
A footnote in the petition reveals the government tried to attribute these discrepancies to Latif changing his story until it became clear the fault lay in the inconsistency of the report writing of one or both of his interrogators.
The government initially argued that the reports were so inconsistent that they proved that Latif, like a guilty man, was changing his story from one interrogation to another. When it was pointed out to the government that the reports were evidently from the same interrogation, and that the discrepancies were created by the government, not by Latif, the government abandoned this argument.
Now, neither of these reports are the report that claims Latif trained with the Taliban, what I suspect is TD-314/00684-02. So showing that the reporting process of that May 29, 2002 interrogation introduced discrepancies is not sufficient to prove that the report at issue suffered from the same–and worse–kind of reporting problems.
But it’s significant to this case that even among the reports not written in the fog of war–as the report at issue was–the reporting process of one or both of these reports introduced (at best) confusion into the report, if not outright inaccuracy. The government, faced with that fact in a case in which they were at the same time insisting that all interrogation reports be accorded the presumption of regularity, simply blamed the detainee and then just dropped it.
If Janice Rogers Brown has her way and such interrogation reports are granted the presumption of regularity, then we must accept that a reporting process that describes Latif as both Bangladeshi and Yemeni, as both as a madrassa graduate and as someone who did not graduate, to have not introduced any inaccuracies.
Now, Henry Kennedy wasn’t pointing to the obvious deficiencies in the Gitmo files when he ruled an interrogation report not credible (though he did suggest those reports might suffer from translation problems, something that several exhibits submitted with the cert petition support). Kennedy was making a much more modest argument: that interrogation reports produced in a process with none of the organization that had been imposed at Gitmo by May 2002 should not be presumed to be accurate records of an interrogation.
If the government can’t even produce consistent reports from a relatively orderly prison, then why has the DC Circuit mandated that courts accept interrogation reports from far more chaotic processes?
One final note: Soufan suggests that if CIA cables have been shown to have inaccuracies, the entire cable is withdrawn. Even Rogers Brown admits that the report in question included an “obvious mistake.” If, as I suspect, this is a CIA cable, and if it has such obvious mistakes that even a Circuit Court judge sees it, then why hasn’t the CIA withdrawn the cable?
Or have they?