Executive Order 14110

Official/legal definition: Executive Order 14110, titled "Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence," is a United States Presidential directive issued October 30, 2023 that sets Federal policy, principles, definitions, and mandates for federal agencies concerning development, evaluation, procurement, and use of artificial intelligence (AI). The Order includes an explicit set of policy principles and a definitions section (Sec. 3) that defines terms such as "artificial intelligence," "AI model," "AI system," "AI red‑teaming," and "dual‑use foundation model," and it directs agencies (including NIST, HHS, DOE, Commerce, and others) to develop guidance, testing resources, labels, red‑teaming practices, and other measures to manage AI-related risks. (See Executive Order 14110, Secs. 1–4; Definitions at Sec. 3). ([govinfo.gov](https://www.govinfo.gov/content/pkg/FR-2023-11-01/pdf/2023-24283.pdf))

Scope and context: EO 14110 established a U.S. federal, risk‑focused approach to AI governance emphasizing safety, security, civil‑rights protections, workforce impacts, consumer protections, and international engagement. It directed technical standardization and testing work (for example, a Generative AI companion to the NIST AI Risk Management Framework and red‑teaming guidance), required agency plans and guidance across sectors (health, labor, energy, education, communications, etc.), and signalled that federal procurement and other levers would be used to encourage compliance with the Order’s principles. The text as published in the Federal Register contains detailed agency tasks and timelines. ([govinfo.gov](https://www.govinfo.gov/content/pkg/FR-2023-11-01/pdf/2023-24283.pdf))

Key requirements and actionable mandates (high‑level):

• Definitions and terminology for federal application (Sec. 3), including "AI model," "AI system," and "dual‑use foundation model." ([govinfo.gov](https://www.govinfo.gov/content/pkg/FR-2023-11-01/pdf/2023-24283.pdf))
• Directives to NIST to produce companion resources (e.g., a Generative AI profile to the AI RMF) and to develop evaluation, red‑teaming, and benchmarking resources. (NIST later published a Generative AI Profile pursuant to these instructions). ([nist.gov](https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence?utm_source=openai))
• Agency‑specific obligations and timelines to address sectoral AI risks (health, labor, energy, national security, biosecurity, etc.). ([govinfo.gov](https://www.govinfo.gov/content/pkg/FR-2023-11-01/pdf/2023-24283.pdf))
• Supply chain and infrastructure reporting and monitoring requirements for large compute/cluster acquisitions and certain IaaS interactions for advanced models (as set out in sections addressing dual‑use foundation models). ([govinfo.gov](https://www.govinfo.gov/content/pkg/FR-2023-11-01/pdf/2023-24283.pdf))

Practical implications for businesses operating across jurisdictions: Businesses that develop, host, or deploy advanced AI models needed to treat EO 14110 as a material policy driver in the U.S. federal context while it was in force: federal contractors faced procurement conditions and assessment expectations (risk management, testing and red‑teaming evidence, documentation for model safety and security). Firms supplying large compute or foundation models saw obligations or expectations to cooperate with federal guidance and reporting. At the same time, non‑U.S. firms selling AI into the U.S. market needed to consider compliance with U.S. agency guidance and reporting obligations tied to the Order (even when their primary regulation would be elsewhere). After the January 2025 change in administration, many EO‑driven agency actions were reviewed or rescinded, which changed the compliance landscape (see "Rescission and current status" below). ([govinfo.gov](https://www.govinfo.gov/content/pkg/FR-2023-11-01/pdf/2023-24283.pdf))

Jurisdictional Variations:

United States (federal): EO 14110 is a U.S. Presidential executive order and therefore directs federal executive‑branch policy and agency action rather than creating standalone statutory obligations. Its definitions and mandates were published in the Federal Register (Executive Order 14110, 88 Fed. Reg. 75191 (Nov. 1, 2023)). Agencies such as NIST implemented companion guidance (e.g., the NIST AI RMF Generative AI Profile) pursuant to the Order’s instructions. Some EO provisions were operationalized through agency guidance, OMB memoranda, or rulemaking initiatives rather than by statute. ([govinfo.gov](https://www.govinfo.gov/content/pkg/FR-2023-11-01/pdf/2023-24283.pdf))

European Union: The EU does not use U.S. executive orders as a source of law. Instead, the EU adopted the Artificial Intelligence Act (Regulation (EU) 2024/1689) which defines AI concepts in Article 3 and creates binding, risk‑based obligations for providers and deployers in the EU internal market. The EU framework focuses on conformity assessment, CE‑style obligations for high‑risk systems, transparency duties for certain systems, and outright prohibitions for some unacceptable uses. EO 14110 does not change or replace obligations under the EU AI Act; multinational firms must therefore comply with both EU regulatory obligations (where applicable) and any U.S. federal agency requirements arising from U.S. policy instruments. (See EU AI Act, Article 3; Regulation (EU) 2024/1689). ([ai-act-service-desk.ec.europa.eu](https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-3?utm_source=openai))

International standards and soft law: International frameworks (OECD Principles, ISO/IEC 22989, UNESCO Recommendation on the Ethics of AI) provide principles, terminology, and ethics guidance but do not create U.S. domestic executive directives. EO 14110 aligned U.S. agency work with existing international standards and asked agencies to promote global standards and cooperation. For example, the OECD Principles articulate value‑based principles (transparency, robustness, accountability), ISO/IEC 22989 supplies technical terminology and concepts used by standards bodies, and UNESCO’s Recommendation sets ethical human‑rights oriented guidance — all of which inform how multinational organizations interpret EO‑driven technical and policy guidance in an international context. ([dig.watch](https://dig.watch/resource/oecdlegal0449-recommendation-council-artificial-intelligence?utm_source=openai))

Rescission and current status: EO 14110 was published in the Federal Register and led to numerous agency actions and guidance between late 2023 and 2024. On January 20, 2025, a new Presidential action (Initial Rescissions of Harmful Executive Orders and Actions, and subsequent AI‑policy EOs) directed review and in many cases rescission or revision of prior EO‑driven actions; and a later order in January 2025 expressly set a new AI policy direction and instructed review of actions taken pursuant to EO 14110. Businesses should therefore check the current, agency‑level status of specific implementing guidance, rulemakings, and procurement conditions rather than assume the EO’s directives remain active in their original form. Authoritative historic text and agency outputs remain relevant for understanding prior expectations and for tracing work products (e.g., NIST publications) that may persist or evolve. ([whitehouse.gov](https://www.whitehouse.gov/presidential-actions/2025/01/initial-rescissions-of-harmful-executive-orders-and-actions/?utm_source=openai))

Examples & cross‑references:

• Example: NIST’s "Artificial Intelligence Risk Management Framework: Generative AI Profile" (NIST AI 600‑1) was developed as a companion resource pursuant to EO 14110 and remains a reference for organizations implementing risk management for generative models. ([nist.gov](https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence?utm_source=openai))
• Related concepts: dual‑use foundation model (definition and reporting expectations in EO 14110); AI red‑teaming; AI RMF / NIST guidance; EU AI Act (Article 3 definitions and risk categories); and international principles (OECD, ISO, UNESCO). ([govinfo.gov](https://www.govinfo.gov/content/pkg/FR-2023-11-01/pdf/2023-24283.pdf))