HomeLegal DirectoryFY22 NDAA §256 (DoD AI T&E)

In effect Limited protection

FY22 NDAA §256 — DoD AI Test and Evaluation Master Plan (Pub. L. 117-81)

United States · Pub. L. No. 117-81, §256 (Dec. 27, 2021)

Section 256 of the FY22 NDAA (Pub. L. 117-81) directed the Department of Defense to develop a comprehensive AI test and evaluation master plan covering data quality, model validation, and adversarial robustness. One of several discrete AI-related provisions across the FY22 NDAA — together they formed the federal government's first comprehensive AI safety-testing framework.

Technical detail

Pub. L. No. 117-81, §256 (Dec. 27, 2021) — directed DoD to develop and submit to Congress a comprehensive AI test and evaluation master plan addressing (1) data quality assurance, (2) model validation and verification methodologies, (3) adversarial robustness testing protocols, (4) deployment monitoring requirements. Related FY22 NDAA AI provisions: §1551 (DoD AI/ML Cyber Threat Information Sharing), §240 (Joint AI Center renamed CDAO), §1542 (AI in cybersecurity).

Who is protected: Public via downstream AI safety-testing standards

Who must comply: Department of Defense

Key facts

JurisdictionUnited States
LevelFederal
StatusIn effect
Protection strengthLimited protection
Effective date2021-12-27
Enacted2021-12-27
CitationPub. L. No. 117-81, §256 (Dec. 27, 2021)
Enforced byDoD Chief Digital and AI Officer (CDAO)
Private right of actionNo — agency enforcement only
Topicsgovernment use of AI
Last verified2026-06-17
Official sourcePub. L. 117-81 (FY22 NDAA, PDF) ↗

More AI rules in United States

Related government use of AI rules elsewhere

See something wrong or out of date? Submit a correction — every entry must carry a verifiable official source.