Federal AI Use Case Inventory

Under Executive Order 13960 (2020) and successor OMB guidance — most recently Memorandum M-25-21 — every covered federal agency must publish an annual inventory of its AI use cases. The Office of Management and Budget consolidates the agency submissions into a single public dataset; this page is a browsable view over that dataset.

Stages

Deployed means the system is in production use. Pilot means it's running with real users in a limited scope. Pre-deployment covers research, prototypes, and acquisitions that haven't yet reached pilot. Retired means the system has been turned off (and is included so the inventory remains a complete historical record).

High-impact

Some use cases are flagged "high-impact" — OMB's term for AI that could meaningfully affect rights, safety, or access to federal services. High-impact systems carry additional governance requirements (impact assessments, ongoing monitoring, public consultation, appeal processes). Toggle "High-impact only" to focus on those.

What's not here

• Intelligence community systems and certain Department of Defense applications, which are exempt from public reporting.
• Use cases an agency lawfully withheld for disclosure-risk reasons. The inventory's is_withheld field flags counts of these without revealing details.
• Common COTS uses (transcription, generic copilots) are reported as agency-level totals in a separate consolidated file and not included here — they would dominate the list without telling you anything specific.
• State, local, tribal, and territorial governments. Federal civilian agencies only.

Refresh cadence

The OMB consolidated inventory is refreshed annually, with mid-year updates from individual agencies. Mungomash rebuilds this page's data when a new consolidation is published. The "Refreshed" date in the hero indicates when this site last synced from the upstream dataset.