Archives Without Borders is a new project of the Columbia Global Policy Initiative. Led by Professors Matthew Connelly, Renato Rocha Souza, and Michael Moss, the project aims to improve the efficiency of government data declassification and to ensure public access to declassified data.
The massive volume of digital documents produced by governments is overwhelming archivists and officials charged with reviewing material for declassification. Without new technologies to prioritize documents that require close scrutiny, less of the official record will be publicly available, reducing opportunities for policy-relevant research and jeopardizing the principal of freedom of information. Archives Without Borders brings together a multidisciplinary team from Brazil, the UK, and the US uniquely suited to address this challenge and advance practical solutions. Fundação Getulio Vargas, a Brazilian social science and information technology institution, houses a large digital collection of official government papers and conducts cutting-edge research on information storage and retrieval. Project Abacá, a partnership between the Universities of Glasgow and Northumbria, conducts research on sensitivity review by experts in British archives and government review processes. The Declassification Engine project at Columbia University has built the largest single repository of declassified US government documents along with natural language processing and machine learning tools to mine this data for new insights about the history and future of world politics. Together these teams will collaborate to produce novel and impactful public policy research, build an international web-based archive, and develop cutting-edge analysis tools to assist declassification review and empower individual researchers.
The project has met with various key policymakers and experts in declassification and document review. At the US National Archives, the team presented new methods that can be used to build a “nuclear classifier” that, given a document, would attempt to say how likely it is to contain “nuclear related information.” The method addresses a specific policy problem that has created a bottleneck in declassification. They also attended the public meeting of the Public Interest Declassification Board—a group jointly appointed by Congress and the President to advise on policy—which has been advocating the use of new technologies in aiding the declassification process. The project will present their research at a future meeting of the Public Interest Declassification Board.
Support from the Catherine T. and John D. MacArthur Foundation and the Columbia Global Policy Initiative