Project Babel · AI Governance Observatory

jurisdiction readout

24 analyzed sources · 2021-07-12 to 2025-08-18

United States' current AI governance record is strategy- and soft-law oriented, with National Institute of Standards and Technology (NIST) and NIST appearing most often in the source base. Across the analyzed documents, Consumer and public safety is the clearest rationale, with Fundamental rights as the next strongest layer, while risk assessment, transparency and disclosure, and human oversight recur most often in the operative expectations.

governance posture

The jurisdiction currently reads as a mixed stack of strategy, guidance, and selective binding measures. The strongest institutional signals come from National Institute of Standards and Technology (NIST), NIST, and NIST Information Technology Laboratory (ITL), and the corpus is weighted toward standards guidance and regulatory guidance.

implementation

Operationally, the sources most often point to risk assessment, transparency and disclosure, and human oversight. Enforcement language is comparatively soft and usually framed through soft enforcement and reporting obligations rather than hard sanctions.

source coverage

This readout is based on 24 analyzed documents spanning 2021-07-12 to 2025-08-18. The corpus is weighted toward standards guidance and regulatory guidance. On the binding side it leans toward voluntary commitment and regulatory guidance, so it captures official policy posture more directly than downstream enforcement practice. The most recent additions in the current mix are Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions and Winning the Race: America's AI Action Plan.

frame distribution

101% of framings observed

281912177333

XRSSECRGTECOINVSOCSAFLAB

framing landscape

Consumer and public safety

Talks about AI as a product-safety issue.

Used when a document explains its rules through consumer or bystander protection — defects, deceptive design, physical or economic harm, duties of care.

Consumer and public safety is the clearest rationale in the corpus, with Fundamental rights and Innovation enablement still materially shaping how governance is justified.

operational profile

recurring requirements and consequences

top safeguard requirements

Risk assessment20 · 83%
Transparency and disclosure18 · 75%
Human oversight17 · 71%
Model documentation15 · 63%

top enforcement hooks

Soft enforcement20 · 83%
Reporting obligations10 · 42%
Regulatory supervision7 · 29%
No explicit enforcement mechanism6 · 25%

key sources in this readout

selected from the analyzed corpus

M-25-21: Accelerating Federal Use of AI through Innovation, Governance, and Public Trust2025-04-03

Regulatory guidanceExecutive actionInnovation enablement

Executive coordination anchor for risk assessment and pre-deployment testing, with reporting obligations as the clearest consequence or oversight hook.

M-25-22: Driving Efficient Acquisition of Artificial Intelligence in Government2025-04-03

Regulatory guidanceExecutive actionEconomic competitiveness

Executive coordination anchor for risk assessment and pre-deployment testing, with procurement conditions as the clearest consequence or oversight hook.

Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions2025-08-18

Regulatory guidanceRegulatory guidanceConsumer and public safety

Regulatory implementation anchor for pre-deployment testing and risk assessment, with regulatory supervision as the clearest consequence or oversight hook.

Joint Statement on Enforcement Efforts Against Discrimination and Bias in Automated Systems2023-04-25

Regulatory guidanceRegulatory guidanceFundamental rights

Regulatory implementation anchor for risk assessment and transparency and disclosure, with regulatory supervision as the clearest consequence or oversight hook.