evidoria

← Back to browse

Good practice Imported

The 2025 Federal AI Use Case Inventory — OMB's Annual Public Ledger of US Government AI

United States of America · Washington, D.C. · See the United States of America profile · See the Washington, D.C. profile

Evidence: Observational / pre–post Top 24% 73/100 · Ask Evidence Copilot about this practice

Under OMB Memorandum M-24-10, US federal agencies must publish a machine-readable annual inventory of every AI use case. The 2025 edition logged 3,611 use cases across 56 agencies — more than double 2024's 1,757 — including 445 classified high-impact.

3,611
AI use cases logged (2025)
56
Federal agencies reporting (2025)
105%
Year-on-year increase in use cases (2024-2025)
1,757
Use cases reported in 2024 (2024)
445
Use cases flagged high-impact (2025)
215
High-impact use cases at Department of Veterans Affairs (2025)
447
Use cases reported by Department of Health and Human Services (2025)
425
Use cases reported by NASA (2025)
367
Use cases reported by Department of Veterans Affairs (2025)
9
Agencies reporting for the first time (2025)

Details

Maturity
Established
Promoter
US Office of Management and Budget (OMB)
Period
2023–present (annual inventory; 2025 edition published April 2026)
Keywords
government, AI governance, transparency, public administration, digital government

Context

OMB Memorandum M-24-10 (2024) requires US federal agencies to publicly post a machine-readable CSV inventory of their AI use cases each year, extending an inventory requirement first introduced under Executive Order 13960. Each entry records a system's purpose, its deployment stage, and whether it is classified as rights- or safety-impacting ("high-impact").

Activities

The 2025 inventory, compiled by OMB and published via GSA's ai.gov and a GitHub repository (last updated 14 April 2026), logged use cases from federal agencies government-wide, recording the flagged high-impact systems separately.

Results

The 2025 inventory logged 3,611 AI use cases from 56 agencies, a 105% increase on the 1,757 use cases reported in 2024. The Department of Health and Human Services (447), NASA (425) and the Department of Veterans Affairs (367) reported the most use cases; 445 were flagged high-impact, with the VA alone accounting for 215 of those.

Conclusions

Much of the year-on-year growth reflects improved reporting compliance rather than solely new AI deployment — nine agencies reported use cases for the first time in 2025 — and the inventory records agencies' self-reported entries rather than independently audited system behaviour, so it measures disclosed AI use rather than verified performance or outcomes.

Implementation
Indicative cost
Low (< €50k)
Time to results
Long (> 3 years)
Staffing & skills
US Office of Management and Budget (OMB) — programme owner, General Services Administration (GSA) — publishes via ai.gov, Agency Chief AI Officers across 56 federal agencies — compliance accountability

Conditions for success

  • Statutory mandate (OMB M-24-10, extending Executive Order 13960) requiring machine-readable annual disclosure
  • Standardised high-impact/rights-impacting flag per use case enabling comparison across agencies
  • Public, machine-readable format (CSV plus GitHub) enabling independent analysis

Common failure modes

  • Much of the year-on-year growth reflects improved reporting compliance rather than solely new AI deployment
  • Inventory records self-reported entries, not independently audited system behaviour or outcomes

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Replication kit

Reusable artefacts from this practice — as published by their sources.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful