Senators Call on Biden to Prioritize Procuring Tech, Staff for Large Scale Declassification
John M Lund Photography Inc/Getty Images
By
Mariam Baksh
Officials at the National Archives say agencies could be using artificial intelligence to more efficiently process a tremendous backlog of appeals.
Sens. Ron Wyden, D-Ore., and Jerry Moran, R-Kan., drew attention to national security implications of failing to declassify epic amounts of data and urged the Biden administration to invest in people and technology to address a backlog of appeals that has continued to pile up since the advent of computers and the internet.
“The failures of the current obsolete system have been extensively documented in reports from the Information Security Oversight Office of the National Archives and the Public Interest Declassification Board, and in public testimony from the Office of the Director of National Intelligence,” the senators wrote in a May letter to President Joe Biden, which they released Thursday. “As the Director of National Intelligence reiterated earlier this year, deficiencies in the current system ‘undermine our national security, as well as critical democratic objectives.’”
The backlog involves way more than just static documents. According to a report to the president prepared by Mark Bradley, director of the ISOO, the digital age has created a never ending flow of data that the government is failing to review and process in a timely manner.
“Thirty years ago, most of this information resided only on paper,” Bradley wrote in the 2018 report. “The government stored it in locked safes and shared it in-person by courier or over secure facsimile lines. Today, the government creates electronic petabytes of classified and controlled unclassified data each month, a deluge that we expect will continue to grow unabated. Digital data now comes in a wide variety of forms, including text, numerical data and graphical images stored virtually in clouds and transmitted by email, instant messages, and chats.”
An Interagency Security Classification Appeals Panel had 1,217 appeals pending review at the end of 2018, according to Bradley’s report, which said that could affect the timely sharing of information with state and local entities, as well as those in the private sector.
Under a 2009 executive order, the PIDB has been proposing solutions, which Bradley and the senators echo, that involve assigning staff within agencies to be dedicated to appropriately classify the information and investing in technology to help with the task.
“The complex interagency process necessary to achieve this long-overdue reform demands active leadership from the White House,” the senators wrote. “We therefore urge that the National Security Council be resourced to undertake this process and that you make clear, to the NSC and across the executive branch, that your administration views the completion of a new or amended executive order as an urgent priority.”
The recommendations suggest artificial intelligence could be employed to categorize and mark the information using associated metadata. Access controls—which also lie at the center of the administration’s ‘zero trust’ approach to cybersecurity—could also be used to appropriately store and transmit the information, according to the officials.
Asserting that classification costs the federal government nearly $18.5 billion a year, Wyden and Moran have proposed legislation that would authorize the director of national intelligence, consulting with other federal agencies, to “set policies, direct resources and promote technical solutions” to address the problem while reporting to Congress.