US AI Executive Order

Understanding the White House Executive Order on AI and how AuditTrue supports organizational alignment.

Disclaimer: This page is for informational purposes only and does not constitute legal advice. AuditTrue is not a law firm. Consult qualified legal professionals for compliance decisions.

Overview

The White House Executive Order on Safe, Secure, and Trustworthy Development of Artificial Intelligence (EO 14110, October 2023) establishes principles for AI safety, security, and equity across federal agencies and industry. It sets standards for safety testing, watermarking, privacy protection, and civil rights safeguards. While primarily directed at federal agencies, the EO creates downstream obligations for AI developers and vendors selling to the federal government, and signals the direction of future US AI regulation.

How AuditTrue Supports Alignment

Safety Testing & Red-Teaming (Section 4)

Section 4 of the EO requires developers of powerful AI systems (foundation models with dual-use potential) to share safety test results and critical information with the federal government. Specifically, the EO invokes the Defense Production Act to mandate reporting for models exceeding defined compute thresholds. AuditTrue's risk assessment engine provides a framework for documenting safety evaluations, red-teaming exercises, and capability assessments — organized and ready for reporting. The platform tracks test methodologies, results, remediation actions, and reporting status in a centralized dashboard.

Infrastructure Security (Section 5)

Section 5 addresses AI's implications for critical infrastructure and cybersecurity. It directs agencies to assess AI risks to critical infrastructure sectors and develop security standards for AI systems used in these domains. AuditTrue helps infrastructure operators document AI system security postures, track vulnerability assessments, and maintain compliance evidence for sector-specific requirements.

Watermarking & Content Authenticity Standards

The EO directs the Commerce Department to develop standards for watermarking AI-generated content and establishing content provenance mechanisms. AuditTrue helps document your content provenance practices, model output labeling procedures, and synthetic content disclosure policies. As standards solidify, AuditTrue's policy templates will be updated to reflect specific technical requirements.

NIST AI Risk Management Framework Alignment

The EO directs federal agencies to use the NIST AI Risk Management Framework (AI RMF 1.0) as the basis for AI risk assessment and mitigation. The AI RMF organizes AI risk management into four functions:

• Govern: Establish organizational governance structures, policies, and accountability for AI. AuditTrue's policy automation module maps directly to the Govern function, providing policy templates, role assignment tracking, and governance committee workflows.
• Map: Identify and document AI system contexts, risks, and impacts. AuditTrue's risk assessment engine performs automated risk mapping across bias, safety, transparency, and accountability dimensions — aligned with NIST's risk categories.
• Measure: Quantify and monitor AI risks using metrics and testing. AuditTrue's monitoring dashboard tracks bias metrics, accuracy scores, drift indicators, and fairness measures in real-time.
• Manage: Implement mitigation controls and respond to identified risks. AuditTrue's risk treatment workflows route identified issues to responsible teams, track mitigation actions, and verify closure.

AuditTrue's compliance mapping tool creates a crosswalk between NIST AI RMF functions and your organization's existing controls, identifying gaps and prioritizing remediation.

OMB Guidance for Federal Agencies

The Office of Management and Budget (OMB) issued guidance (M-24-10) implementing the EO's requirements for federal agency AI use:

• AI governance councils: Each agency must establish an AI governance body. AuditTrue's governance module supports council charters, meeting records, and decision logs.
• AI use case inventory: Agencies must maintain comprehensive inventories of AI use cases. AuditTrue's discovery module inventories AI systems with automated classification by risk and impact.
• Risk assessment requirements: Each AI use case must be assessed for rights-impacting and safety-impacting risks. AuditTrue's risk engine performs automated assessments aligned with OMB risk categories.
• Mandatory minimums: For high-risk use cases, OMB requires specific safeguards — human oversight, real-world testing, independent evaluation, and public transparency. AuditTrue's oversight workflows document each safeguard with evidence trails.
• Waiver documentation: Agencies can waive specific requirements with justification. AuditTrue's waiver module records justifications, approval chains, and expiration dates.

Federal Procurement Requirements

The EO and subsequent federal acquisition rules create specific obligations for AI vendors selling to the federal government:

• Safety test reporting: Vendors of powerful AI systems must report safety test results to the government before deployment.
• Red-team testing evidence: Vendors must document red-teaming methodologies and results. AuditTrue's testing module organizes adversarial test cases, results, and remediation records.
• Performance documentation: Vendors must provide detailed model performance documentation. AuditTrue auto-generates model cards and performance reports aligned with federal requirements.
• Bias and fairness evidence: Federal procurement increasingly requires evidence of bias testing. AuditTrue's bias detection module produces demographic analysis reports and fairness metric summaries.
• Data provenance: Vendors must document training data sources and processing. AuditTrue's data governance module tracks lineage from source to deployed model.

State-Level AI Regulations

While federal AI regulation progresses through the EO and agency guidance, states are enacting their own AI laws:

California (CA)

California has been the most active state in AI legislation. SB 1047 (AI safety) and AB 2013 (AI training data transparency) set requirements for large AI model developers operating in California. The state's privacy law, CPRA, also covers automated decision-making. AuditTrue maps to California requirements with specific compliance templates for CA AI laws.

Colorado (CO)

Colorado's AI Act (SB 24-205) requires developers and deployers of high-risk AI systems to exercise reasonable care to prevent algorithmic discrimination. It mandates impact assessments, risk documentation, and annual reviews. AuditTrue's impact assessment templates and risk documentation workflows align with Colorado's specific requirements.

New York (NY)

New York City's Local Law 144 requires annual bias audits of automated employment decision tools. New York State's SHIELD Act addresses data security for AI systems processing personal data. AuditTrue's bias auditing module produces audit-ready reports aligned with Local Law 144 requirements, including selection rates, impact ratios, and demographic breakdowns.

Other State Developments

Illinois (AI Video Interview Act), Texas (DEI training data law), and Connecticut (data privacy act with AI provisions) have all enacted AI-related legislation. AuditTrue's compliance mapping tool tracks state-by-state requirements and flags obligations based on your operational footprint.

Key Areas We Help Address

• Safety testing documentation — Framework for organizing red-team results and capability assessments
• Bias and fairness auditing — Automated demographic analysis and fairness metrics
• Transparency reporting — Model card generation and capability documentation
• Privacy impact assessment — Data governance and privacy-preserving technique documentation
• Federal agency engagement — Report-ready compliance documentation packages
• NIST AI RMF alignment — Crosswalk between NIST functions and organizational controls
• OMB M-24-10 compliance — Agency AI governance, use case inventory, and risk assessment support
• State law mapping — Compliance templates for CA, CO, NY, and emerging state regulations
• Procurement compliance — Vendor documentation packages for federal AI acquisition requirements

Our Approach

AuditTrue provides the governance infrastructure to organize, document, and monitor your AI systems in alignment with the principles outlined in the Executive Order and subsequent federal guidance. We focus on the technical tools — safety evaluation frameworks, bias detection, and documentation automation — that help organizations demonstrate responsible AI practices. For federal contractors, we help organize the evidence needed to meet procurement requirements. Compliance determinations remain the responsibility of each organization, and organizations should consult with legal counsel regarding specific obligations under the EO and implementing regulations.

See How It Works →  Talk to Our Team