This note is the second in a series which looks at redacted records in the JFK ARC as of February 2025. This new note specifically looks at redacted CIA records, which make up the majority of the 2500+ records that still have some information held back.
The main points of this note:
- About 60 CIA documents redact whole pages
- The majority of CIA redactions have specific, labeled content
- The majority of CIA documents have only one or two redactions
- The subjects of these redacted documents are very miscellaneous
The point of this note is not to argue that it is okay if ARC redactions are not released. The point is to show that, by their very nature, current CIA redactions will not provide major revelations concerning the JFK assassination.
What is a redaction?
“Redacting” refers to removing information (usually text) from a document and replacing it with a “holder” that indicates something has been removed. The ARC uses a blank box as a holder. One blank box = one redaction.
How many redacted CIA records are left?
To understand the overall amount of redaction left in the ARC, first read the previous note in this series, available here.
As that note pointed out, I found that a total of 2544 ARC records still have redactions. Having done such counts several times now, I am well aware that there may be minor mistakes in my counts, but I am confident that this number is very close to correct.
This note will discuss CIA records that still have redactions as of Feb 2025. CIA records use the numerical prefix 104-. By my count, there are 1484 104- records that still have redactions. This is close to 60 percent of the total number of redacted documents.
These records are all listed in the CIA postponement documents discussed in the last post. I will put up a complete list of these records in a more convenient format soon.
CIA redaction categories
The CIA 2022 postponement documents divide information redacted in the records into three basic categories: 1) people-related, 2) location-related, and 3) operational details.
In its letters to President Biden, CIA justified postponing release of this material by claiming its release would injure CIA’s ability to collect intelligence. I will not go into the details of this claim. Many can and have disagreed about such things.
However, these categories do help us sort through the remaining redactions. Each record in the doc index supplied by CIA for its postponement request is marked as one of the categories. We can therefore put similarly labeled docs together and see what the CIA is up to, so to speak.
It is worth noting that this is actually a simplified version of what CIA did much earlier at the bidding of the ARRB. The ARRB wanted just such a system to let people know what was being held back. To do this, they devised a system of “substitute language codes.” There were 34 of these for the CIA, and they provided a fine grained description of “redactable” information.
The current postponement categories simply merge these codes into more general ones. For instance, code 3 was CIA employee, code 4 was CIA asset, code 5 was CIA source, and code 6 was a more general “name of person”. These codes were all merged into the single category of “People.” Similarly, “location” is a simplified category derived from a more detailed set of codes for cable and dispatch locations.
Code 24 was “operational details”, so this redaction category was taken directly from the ARRB system, but it now also includes things like alias documentation (code 31) cover details (code 32), organization names (code 30) and so forth. (Organization name, in the past, was often a fuzzy way of saying liaison services.)
CIA has now marked all of this information for specific documents and specific redactions. This allows us to compile tables for CIA requests. As noted above, there are 1484 redacted records with the 104 prefix. Classifying by the three categories we get:
| category | count |
|---|---|
| People | 385 |
| Locations | 182 |
| Op details | 314 |
| People, Locations | 126 |
| People, Op details | 171 |
| Locations, Op details | 63 |
| People, Locations, Op details | 243 |
This gives us the total of 1484 records. The largest number of redactions are due to people (names). Op details are second, with many of these involving cover and liaison. And there are still a number of records that redact locations. As I wrote earlier, “location” mostly turns out to be the location of CIA stations, as given in cables and dispatches. There are a few safe house addresses and such things, but cable locations are the main thing.
Mixed categories are mostly documents with multiple redactions of different categories of information. After the 2022-2023 releases, many redactions were restored, so there is actually now only one type of redaction in the majority of these docs, rather than two or three. Sometimes it is difficult to decide which category is left, however, so I have simply listed the original CIA categories rather than try to recategorize the records myself.
How much information is still redacted in CIA records?
So we know which records are redacted, and we know what type of information is being redacted, but the volume of material which has been redacted (removed) from these records remains unknown.
One way to estimate this is to look at the size of the redaction boxes. Small box = a single bit of information. Large box = multiple bits of information. If the redaction box is covering something besides text, it’s a different story of course.
It’s quite a pain to get out a ruler and try to measure the physical size of each redaction box left in the collection. Is there a quick and dirty way we can handle this? Well, yes, there is. It turns out there is a correlation between size of box and redaction category.
Redacted locations are almost all cable and dispatch prefixes, as mentioned above. These are each four letters long, so naturally the boxes are all small. (Beware, sometimes a single big box may cover up multiple locations!) Each prefix has a distinct meaning (country and/or city), so in the usual case, each redaction box is indeed a single bit of information.
People redaction boxes are similar. Most people redactions are one or two names. These may be somewhat longer than location prefixes, but not much, so again each redaction box is smaller and again each box usually represents just one bit of information. In addition, most docs discuss just one or two subjects, so that while it looks like a lot of redaction, underlying, it is all just one or two names.
What about op details? Some can of course be very short as well. What type of cover is a CIA officer operating under? This is often just a word or two (army, navy, state, deep commercial, etc.) Other details can be much longer and more involved, however. As a result, these redactions are sometimes longer, a paragraph, or a multi-line text block.
Looking at our category table again, it is certain that not much is held back in the majority of these records, just because the majority of them are people or location redactions.
Another way of estimating how much information is held back in a record is to look at how many redactions are in each document. There are generally many fewer than most people think.
For instance, in currently redacted CIA records, how many have only one to two redactions? The answer: 821. Since the total number of redacted CIA records is 1484, that means about 55% have only one to two redactions. Most of these records are one, two, three pages long, but there are more than a few records that are dozens, hundreds of pages long, with only one or two redactions in them.
Put these two factors together, only one or two boxes, each covering only one name or cable prefix, and that makes a record lightly redacted. It is thus fair to say that the majority of CIA records still redacted in the ARC are very lightly redacted. Redactions are very short, and redactions are very few.
Are there still records which are heavily redacted? Yes, there are, but the number is small. We’ll look at these next.
Whole page redactions
A number of records have what I call whole page redactions. I have discussed these at length. A link to these notes is available here.
In the first release of JFK docs in 2017, there were several hundred records with dozens, sometimes hundreds of whole pages redacted. Most of these pages have since been released, but there are still about 60 records with a total of 600+ pages redacted.
Now whole page redactions does not fit in with the idea of redacting specific bits of information, people, locations, op info, etc. I certainly agree that this type of redaction is not consistent with the low level redactions, targeted at specific bits of information which the ARRB practiced, and which are a basic premise of the JFK Act.
Note, however, that the number of records which have even one such page is very small. How else to describe 60 records in a collection with over 300,000 records? Even this number is semi-inflated. There are only about 20 records with a significant number of whole pages held back.
Most of these pages should be released, or at least opened up so that we can see what specific information needs to be withheld. I will continue posting on these records, since I think the declassification issues here are interesting.
Are there other records that are “heavily redacted” without cutting out whole pages? Yes, there are. First, there are records which are basically just lists where a whole field of information is blanked out. There aren’t many of these docs, but when you see one they stick out, with big blank vertical columns marching down the page.
Some of these are information like social security numbers. SS nums should stay redacted, until the people concerned are deceased. The rest of the list stuff is often book keeping and demonstrably irrelevant to the assassination story. If it’s the names of living people, some of it will be held back for at least a while. But the JFK Act says it must come out, so it will, sooner or later.
There are also a handful of records where there are just a whole bunch of holes, so many that it is impossible to follow what is going on. There were formerly many such docs, but they are almost all gone. One case still around is ARC 104-10439-10113. This is DOI 50-10, a set of regulations which stipulate “Restrictions on Operational Use of Certain Categories of Individuals.” Half of the doc is wide open, the other half is chopped liver.
Why is this doc even in the ARC? Apparently, HSCA wanted to know the details of who the Directorate of Operations can work with, and who is off limits. In my opinion, the parts of these regulations which are still in effect could legitimately be subject to continued postponement.
The miscellaneous nature of current redactions
There are people who find the continued CIA redactions outright suspicious, rather than just a pain in the butt for doing research. This is because they believe that, despite the massive restrictions imposed on CIA redactions, these blank boxes are being used to cover up a story. Oswald was a CIA agent, CIA failed to follow Oswald, CIA was cooperating with the Mafia in the assassination, etc.
In fact, as we have seen, the strict limitations on redaction make it almost impossible to blank out a whole story. What is being covered up then? Names and locations are the easiest to understand.
CIA employee names are redacted when the employee is still under cover, or when the employee retired under cover. This is semi-random; it has nothing to do with the role someone might have played in some assassination related story.
Locations are redacted when a host country objects to release of a station location, or when the station or base location was not declared to the host country. Such cases are also semi-random, and host countries do not object to release of stations based on whether the station appeared in the ARC!
These redactions are thus not related to one story. They cut across stories.
Then there are the files which these redacted docs come from. One of the worst decisions the ARRB made was to take files where the CIA wanted lots of redactions and release them as individual documents. This atomized whole files, turning a single file into dozens of documents, often over a hundred. Tracking files in such cases are a monstrous pain in the ass. For this, blame ARRB, not CIA.
The majority of the records still redacted are single documents from atomized files. They are redacted because of name, location, maybe cover details, maybe mention liaison. It is inevitable that hundreds of files have at least one doc redacted. It is simply impossible to tie all these random bits into one story.
What about records with whole page redactions? There are common types of files, yes. Common subjects? No, I would say.
First, there are personnel files for CIA employees. There are dozens of these with extensive redactions, including over a hundred whole page redactions. Do these employees have anything in common, other than the fact that they were of interest to the HSCA? I don’t see it.
Then there are 201 files for CIA contract agents like June Cobb, Richard Gibson, Henry Lopez, Luis Posada, and WIROGUE. These form another group with many whole page redactions and hundreds of regular redactions. Of interest to HSCA? Yes. Sharing elements of the same story? No way.
Finally, there are CIA projects in Mexico. There are about half a dozen files on these with extensive whole page redactions. These files cover operational details coming years before or years after the assassination.
I’ve looked hard at this set of files. I don’t see a damn thing interesting in what’s open, but there are, of course, blank pages. I wouldn’t bet on anything significant being there, but some people insist on taking the long shot.
Two cents
This is a big picture post. I will back it up with lists, compilations, and examples in the near future.
Again, my point is that CIA doc redactions are not covering up stories, but miscellaneous details These details are inevitably of little or no relevance to the assassination story.
And in my view, even though the details are guaranteed irrelevant, releasing them will hurt. Releasing liaison details, exposing cover mechanisms, outing your own officers are not good things for intelligence collection. Damage is inevitable. The question is, how much damage can you take?
We’re looking at you, Mr. President.