The Central Intelligence Agency maintains a regularly updated electronic archive of declassified historical records that have been publicly disclosed, but it has effectively squandered the utility of digitizing these records by refusing to make them available online.
The CIA to its credit has done more than any other agency to scan declassified records into digital format and to make them word-searchable. Millions of pages of records have been archived in the CIA Records Search Tool (CREST). But those electronic records are maintained in a single geographical location as if they were old-fashioned paper files.
“To use CREST, a researcher must physically be present at the National Archives, College Park, Maryland,” the CIA insists, thereby negating much of the value of the electronic archive.
“Recognizing this presents an obstacle to many researchers, we have been investigating ways to improve researcher knowledge of and access to CREST documents,” the CIA said.
Thus, the CIA web site this week announced a newly updated “finding aid” that permits online word-searches of CREST documents that have been entered into the system through 2008.
But one way to improve researcher access that CIA will not “investigate” is to put the entire CREST database online so that anyone, anywhere could download these declassified, often heavily censored records. Nor will CIA release an electronic copy of the CREST database so that others may post it.
CIA claims that withholding the database from unrestricted release is necessary to protect (what else?) intelligence sources and methods. The Agency evidently believes that there are latent secrets concealed in the declassified record that could somehow be extracted by a clever analyst who reviewed them in electronic form. Further, CIA holds that protecting such ethereal secrets is more important than providing improved public access to the historical record.
It doesn’t matter if this is undemonstrated or untrue. All that matters is that the CIA believes it, or says that it does. But others question the logic of this policy.
“All that the new tool will do,” a classification official in another agency told Secrecy News, “is to drive their FOIA numbers up.”
In other words, researchers who find a document of interest on the CREST search engine but who cannot get to College Park, MD to physically access the CREST archive will tend to request the declassified record under FOIA instead, thereby increasing demand on the already backlogged CIA FOIA process.
So, for example, a search of the CIA CREST finding aid under “Federation of American Scientists” turns up citations to two documents, both from 1972 and both originally of “unknown” classification level: one is a two-page newsletter that was released in full, and the other is an eight-page document that was released in part. The exact nature of either record cannot be ascertained without a trip to the National Archives, or a FOIA request to CIA.
But “this is progress nevertheless,” according to the classification official.
In fact, the CIA CREST database suggests one of the most dramatic and effective enhancements that could be made to the declassification system: Digitize government records as they are declassified and then make them publicly available online.
In 1995, President Clinton ordered agencies that classify information to “establish a Governmentwide database of information that has been declassified” (Executive Order 12958, section 3.8). That never happened, and in 2003 President Bush deleted the requirement (Executive Order 13292, section 3.7). Restoring such a requirement, and fulfilling it, would be an appealing feature of a new executive order on classification.