Melanie Simmons, statistician at DHS/ICE, presents her graduate thesis exploring the national security risks created by the Freedom of Information Act (FOIA) in the age of artificial intelligence. FOIA was designed for a paper-based world, but today, bulk digital data releases can be combined and analyzed by AI to reveal sensitive information never intended for public disclosure. This is known as the Mosaic Theory, and it poses a growing threat to homeland security. Simmons examines how sophisticated actors are exploiting FOIA’s “blind requester” principle, using strategic litigation and large-scale data requests to reconstruct law-enforcement-sensitive databases. She also outlines a path forward: modernizing FOIA to address aggregation and inference risks, integrating existing federal data protection frameworks, and using AI proactively to assess disclosure risks before they become vulnerabilities.
