The Federal AI Use Case Inventory — How Many AI Systems US Agencies Actually Report Using

Executive Order 13960 (December 2020) first required US federal agencies to inventory their AI use cases. OMB Memorandum M-24-10 (March 2024), issued under the White House's 2023 AI executive order, made the exercise mandatory and annual: agencies — with exceptions for the Department of Defense and the Intelligence Community — must inventory each AI use case, submit it to OMB, and publish a public version on the agency's own website.

The 2025 consolidated inventory recorded 3,611 individual AI use cases across 56 submitting agencies, a 105% increase on the 1,757 use cases reported in 2024. Of these, 445 were flagged "high-impact" (triggering the stricter risk-management requirements under M-24-10). Agency-level counts are individually published and auditable: the Department of Health and Human Services grew from 271 to 447 use cases year-on-year, and the Department of Veterans Affairs reported 367 use cases, including 215 high-impact ones.

Because agencies publish their own inventories individually — for example, the Department of Justice's dedicated public AI-inventory page — entries are independently checkable use case by use case, which is a genuine transparency strength. What the inventory does not do is evaluate whether any given use case works well or delivers public value: it is a scale and disclosure metric, not an outcome metric, and the mandate excludes defense and intelligence-community AI entirely, leaving a significant share of federal AI use outside its reach.

Read the full analysis: https://www.whitehouse.gov/wp-content/uploads/2024/03/M-24-10-Advancing-Governance-Innovation-and-Risk-Management-for-Agency-Use-of-Artificial-Intelligence.pdf