This post picks up from the previous one, and continues grinding through comparisons with CIA redactions in 2021 and 2022 in order to get a better idea of what they are still holding out on us. This post looks at “short” (one page) redacted docs. I will suggest below that these short docs are still over-redacted.
As you read this note, remember: there are no CIA documents withheld in full. All the documents discussed below are available online from NARA, with redactions as indicated in the discussion.
From 5600 to 1000 in a year and a half
As we saw in the previous note, there were 5672 one page CIA docs still redacted at the beginning of 2021. After the December 2021 and December 2022 releases, the number is down to 1095.
In a way, however, it’s still a disapppointing result. That’s still a chunk of docs. How many withheld names could be left? How many locations? Did stations really spend all that time talking about liaison relations or mysterious methods in their cables?
Yes, many (most) of these one page records were cables. Cables have always been a difficult hurdle for declassifiers to vault over, with their originating and releasing and authorizing signatures, stations of origin, distribution lists to dozens of other stations, and so on.
This time around CIA cables have indeed been opened up greatly, and in the process have revealed quite a few new names that I didn’t recognize.
William Wainwright was a long awaited name. He was the first case officer for AMLASH-1, previously only known by his pseudonym of “Weatherby”. Or was he? The AMLASH file was like a tidal wave, spawning a huge number of redactions, that are only now slowly ebbing away. With that many docs and redactions floating around, I would not be surprised if Wainwright’s true name had been exposed somewhere, even before this release.
Al Amori was a name I had never heard of before, easily a few hundred documents with his name now revealed on them, mostly Cuban related, as we might expect. I have not yet had a chance to look at his positions, which were multiple.
In general, although they have been tidied up quite a bit, cables still remain a significant component of redacted records, 500+ of the one pagers are still cables, half of the total number of redacted one-page docs as of December 2022.
Redactions per page
At the beginning of 2021, there were thousands of redacted records which held a grand total of one redaction on one page. As a result, CIA showed up in the 2021 database update with 10,000 plus records still redacted.
This incredibly inflated number was the result of two things: 1) a rigid set of guidelines as to what should be redacted; 2) very inconsistent reviewing. The 2022 release has swept a lot of these problems away, but in the end 363 records, a third of all the one-page records, still have just one redaction. That means that these 363 records were released in 2022 without any change at all. A disappointment, for sure. I suspect some of these redactions are still not justified.
I did a table of the number of redactions per page for the short docs in 2021. What would it look like now? Well, if you ignore the one redaction docs, which are, obviously, unchanged from last time, you are talking about a lot more eye-blurring scanning for boxes. In fact, many of these still redacted docs have actually seen some releases, e.g. two redaction docs have become one redaction docs, 3 redaction docs have released one or two redactions, and so on.
No doc has added redactions (that I have seen yet). So we can at least get an upper limit by simply using my original 2021 count. Eventually I will have to count redactions in the new releases (obsessive-compulsive, the doctor called it). But I won’t drag you along with me for this post. Instead, using my old figures, here is a maximal estimate of redactions left in the short docs, compared with 2021:
| # of redacts | # of docs (2022) | percent (2021) | percent (2022) |
|---|---|---|---|
| 1 | 363 | 49% | 33% |
| 2 | 233 | 23% | 21% |
| 3 | 150 | 14% | 13% |
| 4 | 101 | 6% | 10% |
| 5 | 85 | 3% | 8% |
| 6 | 51 | 2% | 5% |
| 7 | 38 | 1% | 3% |
In addition, there are 74 records with more than 7 redactions apiece, about 7% of the total. Remember, this is just a quick, dirty approximation, I have not actually gone through and counted total redactions for each doc in the new releases.
Justifications
One of the more interesting aspects of the 2022 release is that CIA’s document index specifies which of three categories of information is implicated in the document: people, locations, and operational details. Taking the simplest case, 1 page 1 redaction docs, the remaining redactions stack up as follows:
| reason | # of docs | percent of total |
|---|---|---|
| people | 196 | 55% |
| operations | 109 | 30% |
| locations | 41 | 11% |
| mixed | 15 | 4% |
Again we see that people’s names are the main thing being held back here.
My two cents
Given the age of some of these documents, CIA should give careful consideration to withholding so many names. The vast majority of these names are employees, not vulnerable assets or agents. I am reasonably sure that this distribution holds true across all the remaining redacted records. People’s names continue to be the main component redacted. This must be addressed.
That locations make up 10% of the total also surprises me. Locations were a big problem in getting redactions released, and a major source of inconsistencies in redactions. I can tell just by looking at location codes that there is still a problem here. 16-26 is San Salvador. Why redact the name and leave the code? Not acceptable.
In sum, while accepting that there are operational considerations and employee dilemmas, there is clearly still room for more redactions.