Triangulating Across U.S. Federal AI Transparency Regimes

Fragmented disclosure creates dangerous gaps: The U.S. government uses three separate transparency systems (SORNs, ICRs, and the AI Use Case Inventory) to disclose federal AI use, but none individually reveals how government AI systems are actually built or deployed—meaning the public cannot get a complete picture of AI systems that make consequential decisions about benefits and deportation.

Systems can operate in secret for months: Federal agencies can deploy AI systems months before they appear in any official public record, due to the annual cycle of the AI Use Case Inventory, creating a significant lag between when AI systems are actually used and when the public learns about them.

Current records lack basic tracking tools: The disclosure infrastructure lacks persistent identifiers and has widely varying levels of detail, making it extremely difficult for the public to follow specific AI systems across different regulatory regimes or track how they change over time.

Linking records helps but isn't enough: Even when researchers manually connect disclosures across all three regimes, the combined information still falls short of what journalists and researchers have independently discovered about the same government AI systems, indicating structural inadequacy in current oversight.

Transparency gaps are by design, not accident: These disclosure weaknesses stem from each regime's original administrative purpose rather than being fixable oversights, suggesting that meaningful reform requires rethinking how federal AI transparency is fundamentally structured.