This note takes a brief look at the system CIA is now using for JFK records declassification. I am posting this note since declassification is one of my interests; I also hope this note will help clarify some ARC release issues.
The MORI system
This section is based mostly on a document from the National Archives titled “Records Management in the Central Intelligence Agency”, available from the Federal of American Scientists website here.
MORI stands for “Management of Officially Released Information”, and is an integrated system for information management. Development of the system began at CIA in 1994. Not surprisingly, the JFK Act was an impetus for the development of MORI.
MORI replaced two earlier systems, ORIS (Officially Released Information System) and IPS (Information Privacy System. In addition to JFK assassination records, it includes “documents relating to CIA actions in Guatemala, the “October Surprise” conspiracy theory, and U.S. troop exposure to toxic agents during the Gulf War.
MORI went online in 1998, and MORI docids can be found on many declassified CIA documents available today.
CADRE
MORI was not intended as a permanent solution to maintaining records of CIA declassification, and by 2001, another system was in the works: CIA Automated Declassification Review Environment (CADRE).
Although I have not found the kind of systematic introduction to CADRE found in NARA’s discussion of MORI, there is a 275-page pdf on the CIA Records Declassification Program (RDP) available from the Government Attic website (here) which includes a fair amount of information about CADRE.
This doc shows that the first pilot project for CADRE was in 2005, and that by 2016 a “next gen” CADRE was in the works. The document ends in 2016, so no idea what the current status is. There is still a “cadre reference id” used in CIA releases as of 2022, however, so some variant of the name is still in use.
What is the difference between MORI and CADRE? Most notably for the JFKARC, CADRE includes DDR: duplicate detection and resolution. This feature was probably responsible for some confusing “additions” to the ARC in the post-2021 releases.
MORI and CADRE in the JFKARC release process
The Government Attic pdf on the RDP has a couple of reference to JFK records. In a January 2000 report, for example, we learn that the JFK collection (job number CSI-1999-0021) continues to be processed through the “Document Conversion Center” (one of the components of the CIA declassification “factory”). “Currently, more than [redacted] pages have been entered into the MORI database. A total of [redacted] boxes of prepped material has been moved to MORI; [redacted] boxes are “in process” (scanning, indexing, QA) and [redacted] are in hierarchy.”
In a January 2010 report, under the CADRE support summary, we learn that CIA “conducted two rounds of validation on updates to improperly migrated JFK documents, working with the team member who updated the records to ensure that the documents were properly connected by CADRE Reference ID and requesting that the data validation lists be regenerated to account for the newly updated documents. Researched missing release dates on JFK documents in the CADRE MORIPROD database and sent a note to the Migration Team to request feedback on the plan to get this data migrated to CADRE.”
Here we learn of CADRE reference ids being used for ARC documents. These do not show up in the NARA info until much later, ca. 2022.
In a January 2011 report, we read that the RDP “continued to coordinate with members of the JFK migration/validation team to plan a status meeting and to ensure progress on the effort to identify gaps in the data fields required to process JFK documents in CADRE.”
By this point, according to the next entry in the report, CIA was working on CADRE release 2.5.1.
From these reports we can see that CADRE has been in use in the JFKARC for quite some time. The first time I saw an indirect reference to the system in NARA documentation, however, was in 2022.
The CIA postponement docs include a lengthy list of CIA records in the ARC, titled “CIA Document Index”. The list identifies which records were released in full in 2022, and which records were still released only in part (i.e., they still had redacted text as of December 15, 2022). The list is available here
Column headings in the doc include “Subject”, “Case number”, “CADREREFID”, “CIA/OGA prefix”, “NARA ID”, “pub date”, “decision”, “# of pages”, “justification”, and “proposed date for rereview”. We can now understand all of these.
The subject is the “re-review” of JFK Act documents. The case numbers for the 2022 releases ranged from “SC-2019-10068” to “SC-2022-00011”. These are tracking numbers for the releases, created for the Records Declassification Program. CADREREFID is the cadre reference id number, already mentioned above. Each of the 11533 records in this CIA list has a unique cadre ref ID. The CIA/OGA prefix is the agency prefix in the JFKARC RIF numbers, OGA standing for other government agency. The meaning of the other headings is obvious.
The cadre ref ids are easily identifiable: each starts with the letter “C” followed by eight digits. For these records, the first two digits are always ’00’. The “cad refs”, as I’ll call them, are not just for bean counting. Beginning in 2022, there are several records where the cad ref is suffixed to NARA’s RIF number to make a unique filename. This happens with ARC 104-10326-10014, which is suddenly split into two files: 104-10326-10014[c00564689].pdf and 104-10326-10014[c06931192].pdf
Both of these pdfs seem to be copies of the same cable, but a close look shows that cable item 7 differs in the two copies. This is an example of CADRE’s ability to “detect duplicates” in the sense that what was originally treated as one record turned out to be two distinct variants.
This is a problem, because the NARA RIF number is supposed to be unique for each JFK record. CADRE detection thus requires subversion numbers in the numbering system. Major hassle: it took me quite a while to figure out how to handle this. Not something my teachers discussed in graduate seminars on Han dynasty syntax.
There are other examples of CADRE suffixes which show up in the 2023/2025 pdf filenames; moreover, this filenaming convention is not limited to CIA owned records.
JFK records supplied by CIA have an agency prefix of 104. However, CIA is also responsible for some of the classified information in other agencies records. These records were also put into the CADRE pipeline, and they also have cad refs. An example of this shows up in the 2022 pdfs of FBI records, where we find 124-10167-10378[c06716615].pdf, 124-10167-10378[c06716617].pdf, and 124-10167-10378[c06716618].pdf.
In fact, this analysis was used in the ARC as far back as 2018. These same three FBI records show up there as 124-10167-10378.pdf, 124-10167-10378_1.pdf, and 124-10167-10378_2.pdf. The difference is simply in the filenaming convention; in 2018, they used file counts to keep the filenames distinct, whereas in 2022, they started using the actual cad ref.
This issue may account for some filename problems in 2017 as well. This is perhaps trivia, but it has caused very annoying years-long problems for ARC bean counters like me.
It is worth noting that this helps explain at least some of the “file fragments” I discussed in a series of notes starting here. As I suggested in the series, these “fragments” are a product of CIA record review. We can now say for sure that this is the CADRE system at work. Too bad NARA didn’t bother to tell us. Of course I didn’t ask them either.