WASHINGTON, March 15, 2024 – Sunshine Week is an annual nonpartisan celebration of the importance of publicly available records and the practice of open government to drive civic engagement. In recognition of the week, on March 14, the National Archives and Records Administration (NARA) hosted a panel discussion titled, “Artificial Intelligence: The Intersection of Public Access and Open Government.”
Director of the Office of Government Information Services Alina Semo kicked off the event and introduced Deputy Archivist of the United States William Bosanko.
“For me, personally, today has been many years in the making,” said Bosanko. “I’ve spent my entire NARA career, more than 30 years, focused on access to records. . . . While we at the National Archives continue to face the seemingly endless challenges and opportunities for our work that I have seen in the last three decades, it is clear that the next frontier is artificial intelligence [AI] and machine learning. AI will undoubtedly change the way the government does business.”
The panel included NARA Chief Innovation Officer Pamela Wright, serving as moderator, in conversation with NARA Chief Data Officer Gulam Shakir; Library of Congress Senior Innovation Specialist, Digital Innovation Lab,
Abigail Potter; Deputy Assistant Secretary for the Office of Global Information Services at the U.S. Department of State Eric Stein; and Director of the Office of Information Policy of the U.S. Department of Justice Bobak Talebian.
The group was selected for their expertise in open government operations and shared transparency values regarding artificial intelligence and how it intersects with the Freedom of Information Act (FOIA) and access to government information.
“The pillars of open government are transparency, collaboration, and participation,” Wright said. “The work we do through our open government plan helps us to hold our government accountable and instills trust with the American people. I cannot think of a more important mission for our work as federal employees in public service.”
The panelists offered a variety of perspectives and current frameworks, based on their respective federal agencies, on record digitization and the technology available to process and record metadata to make records more searchable for a wider audience.
“There are programs and high-level assessments we want to run,” Potter said. “There are things we want to track around fairness and balance.”
Visit the Library of Congress’s AI framework website to view the work being done.
AI is also being used to declassify applicable records, increasing public access. The convergence of the technology to readily and efficiently cull data from scanned records, while still protecting personally identifiable information (PII), aims to provide greater access to those searching for government records. By combining these technologies, the value of agency records to the nation is able to grow.
“This work is done with the intention of making our FOIA process more efficient,” Stein said. “AI and machine learning are very valuable [for] saving time and energy by not having to do the same tasks over and over again.”
Stein then walked the audience through the utilization of AI machine learning in the State Department’s Virtual Reading Room and how the AI search mechanism can expedite identifying applicable records for requesters.
“Going forward, we want to work with agencies and issue guidance to help standardize and make more accessible FOIA logs and frequently requested records so that they can more in real time be digested,” Talebian said during his presentation on how foia.gov is utilizing their new FOIA Wizard tool, which uses AI to direct users to the correct agency for their requests.
Shakir followed with an in-depth presentation on the various projects NARA is spearheading on the integration of AI and machine learning with digital holdings from the consolidation of classified Presidential records to the National Declassification Center to improved searches on the NextGen Catalog.
“We have close to 300 million documents in the National Archives Catalog. So, we are trying to target how we can improve the forward-facing results there,” Shakir said. “Traditionally, you want people to, whatever they are finding, to find it on the first page. So, we are trying out different OCR [optical character recognition] capabilities and metadata extraction to improve the results. PII detection is another bucket where we found we are doing a lot of public release of documents. We are aggressively trying several different out-of-the-box tools.”
A question-and-answer session followed the panel discussion.
View the panel discussion on the National Archives YouTube channel.
For more information on Sunshine Week, visit the National Archives website.