EveryAILaw.comDoes EO 14319 — Preventing Woke AI in the Federal Government require Data Governance?
United States • enforcing
Yes — 1 provision
Requirements at a glance
This regulation imposes 3 specific requirements for Data Governance across 1 provision:
- Data provenance — Training data sources and provenance must be documented and disclosed to procuring agency
- Risk documentation — Vendors must document risks, limitations, and mitigations for LLM systems
- Compliance verification — Agencies must have sufficient documentation to assess vendor compliance with Unbiased AI Principles
Training Data and Risk Governance for Federal AI
Vendors must disclose training data provenance, limitations, and risk mitigations as a condition of federal procurement. While framed as ensuring "unbiased AI," the practical effect is a data governance disclosure requirement for the federal AI supply chain.
Requirements
| Requirement | Details |
|---|---|
| Data provenance | Training data sources and provenance must be documented and disclosed to procuring agency |
| Risk documentation | Vendors must document risks, limitations, and mitigations for LLM systems |
| Compliance verification | Agencies must have sufficient documentation to assess vendor compliance with Unbiased AI Principles |
Penalties
| Violation | Fine |
|---|---|
| Non-compliance | Contract termination; vendors accountable for certain costs |
Sources: EO 14319 Text OMB M-26-04