Responding to a February 21 New York Times story indicating that thousands of declassified documents had been reclassified by executive branch agencies and removed from public access in questionable circumstances, the National Archives and Records Administration (NARA) announced yesterday that an official investigation into the matter was underway.
An audit is being conducted by the Information Security Oversight Office, a NARA component, to determine the number of documents that have been withdrawn, the authorization and justification for the withdrawal, and the appropriateness of the reclassification action.
(Agencies dispute that any documents have been “reclassified.” Instead, they contend, the withdrawn records were never properly declassified and so have remained classified all along.)
“The audit will result in a public report designed to provide the greatest feasible degree of transparency to this classification activity,” NARA said. “It is anticipated that the report will be available within the next 60 days.”
See “The National Archives Responds to Reclassification of Documents,” NARA news release, February 22.
The reclassification issue is more than a minor bureaucratic glitch. It has become a threat to the integrity of the entire national security classification and declassification program.
It would not be surprising if there were isolated cases of mistaken declassification. But because many of the now-withdrawn documents are widely available in the public domain, on the National Security Archive web site and elsewhere, anyone can see that the authority to reclassify and remove them has been improperly exercised in many cases. Government officials have admitted as much.
“If those sample records [reviewed by the Information Security Oversight Office] were removed because somebody thought they were classified, I’m shocked and disappointed,” ISOO Director Bill Leonard told the New York Times. “It just boggles the mind.”
But if records were mistakenly withdrawn in this case, what confidence can anyone have that classification authority is being properly invoked, for example, in the ongoing review of declassified historical records conducted by the Department of Energy? That review, conducted under the 1999 Kyl/Lott amendment, has also led to the removal of many thousands of DOE and other agency records that supposedly contain classified nuclear weapons information.
And what about Bush Administration classification of present-day records, which has accelerated to a record high level? How credible are those classification actions?
Finally, what is the role of the National Archives? Is NARA the guardian of public access to historical records? Or has it become a passive accomplice to the classification abuses of other agencies?
Rep. Christopher Shays (R-CT) announced that his House Government Reform subcommittee on National Security will hold a third hearing on classification policy on March 14.
Commercial artificial intelligence tools have recently emerged that are able to produce police reports. If the resulting reports are inaccurate, incomplete or biased, or if the process leaks confidential information, this could undermine the criminal justice system and harm citizens.
Too often, affected patients, clinicians, and regulators cannot see how the system works, why a decision was made, or whether meaningful human oversight occurred.
Existing tools from other domains, such as existing robust public engagement processes in drug development, when applied to AI deployment can help strengthen public trust in these systems and enhance perceptions of their legitimacy and the decisions they produce.
With thoughtful policy action, it is still possible to build systems that are fair, transparent, and accountable, and to earn the public trust that will ultimately determine AI’s future. We hope policymakers are ready to act.