When President-Elect Joe Biden takes office at noon on January 20, 2021, Executive Order 13526 (E.O.) will be over 10 years old. The E.O. “prescribes a uniform system for classifying, safeguarding, and declassifying national security information, including information relating to defense against transnational terrorism”— but a lot has changed over the past decade. Transnational terrorism now shares a place among many emerging threats to democratic order that include the propagation of disinformation, and cybersecurity. Government operations now require rapid information sharing and secure communications. Since E.O. 13526 came into force in 2009, there have been powerful innovations in information technology, an unending proliferation of electronic data, and means to communicate.
In its May 2020 report to the President, the Public Interest Declassification Board (PIDB) called for Government-wide reform of the nation’s outdated classification and declassification system.
Overwhelmed by the deluge of electronic national security information, declassification remains paper-based and disconnected between many Executive branch agencies that lack such basic tools as classified network connectivity to perform interagency equity reviews. The PIDB’s 2020 report followed seven reports to the White House or the President since 2008 that recognized the still accelerating consequences of digitization and the urgent need to apply innovations such as artificial intelligence, machine learning, and robotic process automation to modernize declassification and reform information security.
Yet, the challenge remains and expands.
With a sense of urgency, the PIDB wants your ideas on how best to modernize classification, declassification and information security policies and practices, including how to use technologies to accomplish those objectives. Please submit your ideas as comments to this blog, or as emails to the PIDB email address: pidb@nara.gov.
The PIDB and staff hope to use your ideas and comments as it begins its outreach and advocacy efforts with the new Administration and the new Congress.
This statement is not an official statement or recommendation from the Department of the Air Force or the Air Force Declassification Office. This is just the opinion of an access professional that has been conducting declassification review for 23 years and is one of the very fewpeople that delivers comprehensive declassification training to agencies across the Federal Government.
The most effective way to expedite the review process is to hire/staff agencies with well-trained, well-compensated personnel to the fullest extent practicable. This will result in the greatest release of properly reviewed documents possible.
While computerized review gives the impression it will expedite review, possibly replacing human review, it does not. I have participated in investigations of various artificial intelligence programs over the last 15 years. So far, no system has been found that is capable of conducting review as fast or as accurately as a well-trained Government Information Specialist. The USAF as well as other Government agencies have looked into this type of review, spent millions of dollars, and are no closer than they were 10 years ago. A company that has worked with the AFDO claims they have a system that is cost effective and reviews faster than a human; however, that is false. The product is a “dirty word” search at best. It is in no way capable of conducting declassification review at this time. The USAF has over 400 Classification guides, that is just the Air Force. To create an all-encompassing program to review all Government records, you would need to incorporate thousands of Security Classification Guides from every federal agency which could have classified equities in a record, along with Technical manuals and other program/policy guides. The system would have to be able to apply those guides/instructions to millions of pages of information.
From 2008 to 2015 our office, AFDO, reviewed/processed about 67,000,000 pages. This does not include records reviewed at NARA’s Joint Referral Center, or FOIA or MDR cases. That is about 9.7 million pages a year, the overwhelming majority being declassified. A knowledgeable reviewer can examine between 600,000 and 1,000,000 pages of information a year IAW EO 13526, the Atomic Energy Act and all applicable government regulations. About 200,000 pages a year when conducting FIOA (Classified) or Mandatory Declassification Requests as these reviews require a line-by-line assessment, redactions, justifications and administrative actions.
The training timeline for a qualified Government Information Specialist is between 6-9 months. By the end of the training cycle, they can review records IAW EO 13526, the Atomic Energy Act, and apply any exemption or exclusion. All of the documents reviewed by outside vendors needed a secondary review by a person. We have invested time, money and a lot effort in several projects over the years, as have other federal agencies, and we’ve seen nothing that can take the human out of the reviewing loop. Sometimes, the automation puts an extra burden on the human reviewer. At best the experiment could be used to prioritize review. The program could rate documents into separate categories: Highly likely to contain classified information, likely to contain classified information, and not likely to contain classified information. This could help speed up the review process, but it doesn’t negate the need for a human review. That process would need to have a “Return on Investment” conducted before any implementation.
Many Government agencies have looked into computerized review; however, none exclusively use it to conduct declassification review. The best way to get these records out to the public is to produce extensive training programs, populate declassification activities with qualified reviewers, and prioritize records for review. Efficient declassification review is accomplished by experienced reviewers with significant time in their organization, who remain with their organization through appropriate compensation and by holding them in high regard. If agencies were able to add and retain more Government Information Specialists to their team, it would produce millions of additional pages released to the public each year.
If you have any questions, comments or would like or have a discussion about these views I can be reached at jhunter163@msn.com. I am also willing to join the panel at any time, in an unofficial capacity.
V/R,
Jeff Hunter