---
id: GOV-US-001
title: "United States AI Policy"
alternative_names:
  - "US AI Governance"
  - "Federal AI Policy (United States)"
category: "Safety & Governance → Regulation"
volume: VIII
chapter: 48
entry_type: "Standard Article"
level: 3
status: "Draft"
last_reviewed: 2026-10-18
version: "0.2.0"
prerequisites:
  - "GOV-LAW-001"
related_entries:
  - "GOV-EU-001"
  - "GOV-CN-001"
  - "GOV-EXPORT-001"
  - "GOV-INTL-001"
  - "ORG-SAFETY-001"
  - "ORG-STANDARDS-001"
---

# United States AI Policy

*Legal and policy status as of 2026-10-18 (fact-check pass; see the audit for what was and was not re-verified). Executive policy in the United States changes by order; re-check primary sources before relying on any statement here.*

## 1. Concise Definition
United States AI policy is the combination of executive orders, agency guidance, existing sector laws, federal funding and procurement rules, export controls, and state legislation that governs the development and use of artificial intelligence in the US. There is, as of 2026-10-18, no comprehensive federal AI statute located in the sources reviewed comparable to the EU AI Act ([[GOV-EU-001]](../vol08_safety_governance/GOV-EU-001_the_eu_ai_act.md)).

## 2. Overview & Significance
The US hosts many of the frontier AI developers, so its policy choices have effects beyond its borders. Policy since January 2025 has emphasised removing perceived barriers to AI development and international competitiveness, in contrast with the 2023-2024 emphasis on safety reporting. Because most rules are executive actions or agency practice rather than statute, they can be, and have been, reversed quickly. The state level has become a second centre of rule-making, which the federal executive branch has sought to limit.

## 3. Historical Development
**Pre-2023.** Federal action consisted of strategy documents and funding (for example, the National Artificial Intelligence Initiative Act of 2020, Division E of Public Law 116-283, dated 1 January 2021 (GovInfo, 2021)) and voluntary guidance. The National Institute of Standards and Technology (NIST) published the AI Risk Management Framework 1.0 (NIST AI 100-1) in January 2023 as a voluntary framework (NIST, 2023).

**Executive Order 14110 (30 October 2023)**, "Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence," required, among other things, reporting to the Commerce Department by companies developing dual-use foundation models, using Defense Production Act authority (Federal Register, 2023). It was **revoked** by Executive Order 14148 of 20 January 2025, "Initial Rescissions of Harmful Executive Orders and Actions," whose Section 2 lists Executive Order 14110 among the revoked orders (Federal Register, 2025a; 90 FR 8237, 28 January 2025).

**Executive Order 14179 (23 January 2025)**, "Removing Barriers to American Leadership in Artificial Intelligence," states a policy "to sustain and enhance America's global AI dominance", directs an AI action plan within 180 days, and directs an immediate review of actions taken under the revoked Order 14110 (Federal Register, 2025b; 90 FR 8741, 31 January 2025). The order itself describes Order 14110 as "revoked" in its Section 5.

**America's AI Action Plan (23 July 2025).** The plan, titled "Winning the Race," is organised around three pillars, which its table of contents names "Accelerate AI Innovation," "Build American AI Infrastructure" and "Lead in International AI Diplomacy and Security"; the plan text is dated July 2025 and the White House announcement page is dated 23 July 2025 (The White House, 2025d, 2025e). It is a policy document, not self-executing law.

**Standards body rename (June 2025, and again by October 2026).** Trade press reported that the Commerce Department announced in early June 2025 that the US AI Safety Institute, created in 2023, would become the Center for AI Standards and Innovation (CAISI) within NIST, with a stated focus on demonstrable risks such as cybersecurity, biosecurity and chemical weapons; NIST news items from 2025 and early 2026 use the CAISI name. As read on 2026-10-18, NIST's own page for the body (nist.gov/caissi, last modified 3 September 2026) names it the Center for Advancing Innovation and Standards for Super Intelligence (CAISSI), and describes it as industry's primary point of contact in the US government for testing and collaborative research on commercial "SI" systems (NIST, n.d.); a NIST news item dated 16 October 2026, issued with the National Security Agency and CISA, uses the same name (NIST, 2026). This article uses "CAISI" for the body under either name. [VERIFICATION NEEDED: the Commerce press release itself could not be retrieved (commerce.gov blocks automated access), so the exact date of the June 2025 announcement is unconfirmed from a primary source, and no announcement of the CAISSI rename or its date was found; see [[ORG-SAFETY-001]](../vol10_people_organizations/ORG-SAFETY-001_ai_safety_institutes_and_research_organizations.md)].

**Executive Order 14365 (11 December 2025)**, "Ensuring a National Policy Framework for Artificial Intelligence," published at 90 FR 58499 on 16 December 2025 (Federal Register, 2025c). Its text directs: the Attorney General to establish, within 30 days, an AI Litigation Task Force to challenge State AI laws inconsistent with the order's policy, including on interstate-commerce, preemption or other grounds (Sec. 3); the Commerce Secretary to publish within 90 days an evaluation identifying onerous State AI laws (Sec. 4); Commerce, through NTIA, to issue within 90 days a BEAD Policy Notice making States with such laws ineligible for non-deployment BEAD funds "to the maximum extent allowed by Federal law," and agencies to assess conditioning discretionary grants (Sec. 5); the FCC Chairman to initiate a proceeding on a federal reporting and disclosure standard for AI models that preempts conflicting State laws (Sec. 6); the FTC Chairman to issue a policy statement on when State laws requiring alteration of truthful AI outputs are preempted by the FTC Act (Sec. 7); and the preparation of a legislative recommendation for a uniform federal framework preempting conflicting State AI laws, which is not to propose preempting otherwise lawful State laws on child safety, AI compute and data-centre infrastructure, or State government procurement and use of AI (Sec. 8). Like Order 14179, it creates no private rights (Sec. 9(c)). The White House fact sheet of the same date summarises these directives (The White House, 2025f). Whether the Sec. 4 evaluation, Sec. 5 notice, Sec. 6 proceeding or Sec. 7 statement were issued was not checked [VERIFICATION NEEDED].

**National Policy Framework for AI (20 March 2026).** A set of non-binding legislative recommendations to Congress, supporting broad preemption of state AI laws imposing "undue burdens", no new federal AI regulator, and preserving state authority over generally applicable laws such as child protection (reported by WilmerHale, 2026, which dates the release 20 March 2026 and describes it as not binding and as advising Congress not to create a new federal AI regulator; primary White House text not retrieved [VERIFICATION NEEDED]).

## 4. Technical / Conceptual Explanation
**Instruments in use.**
- *Executive orders* bind agencies and federal procurement but do not create private rights and can be rescinded by a later order.
- *Existing statutes applied to AI*: consumer protection (FTC Act), civil rights and anti-discrimination law, product liability, copyright, sectoral regulators (FDA, SEC, CFPB).
- *Voluntary standards*: the NIST AI RMF and CAISI evaluations.
- *Export controls* on advanced chips and model-related items are covered in [[GOV-EXPORT-001]](../vol08_safety_governance/GOV-EXPORT-001_export_controls_and_ai_geopolitics.md).
- *Federal funding conditions* as a lever over states (Executive Order 14365, Sec. 5).

**Compute thresholds.** Until Commerce defined technical conditions, Order 14110 Sec. 4.2(b) required reporting for models trained using more than $10^{26}$ integer or floating-point operations (more than $10^{23}$ for models trained primarily on biological sequence data) (Federal Register, 2023); with the order's revocation, that federal reporting requirement is no longer in force. California's SB 53 is reported by a law firm to define "frontier model" as one trained with more than $10^{26}$ operations (below).

## 5. Types, Components & Classifications
- **Federal executive branch**: White House Office of Science and Technology Policy and National Security Council; Commerce (NIST/CAISI, Bureau of Industry and Security); Justice (the litigation task force); FTC and FCC.
- **Congress**: numerous bills introduced; as of mid-October 2026 secondary sources report that no federal preemption statute has been enacted. StateScoop reported that the Senate voted 99-1 on 1 July 2025 to remove a proposed state-law AI moratorium from the budget bill (StateScoop, 2025a), and, on 8 December 2025, that the text of the 2026 National Defense Authorization Act did not include such a moratorium (StateScoop, 2025b). The status of 2026 bills (reported as a "Great American AI Act" draft and a "FRONTIER Act") is described inconsistently in secondary sources [VERIFICATION NEEDED: check congress.gov].
- **States**: California's Transparency in Frontier Artificial Intelligence Act (SB 53) was signed on 29 September 2025, according to the Governor's office (Office of the Governor of California, 2025), which states that it requires large frontier developers to publish a framework on their website, creates a mechanism to report critical safety incidents, protects whistleblowers and sets civil penalties. Law-firm summaries add a 1 January 2026 effective date for key provisions and a more-than-$10^{26}$-operation threshold for "frontier models" [VERIFICATION NEEDED: the chaptered bill text could not be retrieved (leginfo.legislature.ca.gov returned a bot challenge), so the threshold, effective date and penalty amounts are from law-firm sources only]. Colorado: the Legislature's bill page shows SB 26-189 ("Automated Decision-Making Technology") signed 14 May 2026 and repealing and re-enacting the provisions of the 2024 Colorado AI Act (SB 24-205), with some duties starting 1 January 2027 (Colorado General Assembly, 2026). A law-firm alert reports that xAI sued over SB 24-205 on 9 April 2026, that the US Department of Justice filed a companion complaint on 24 April 2026, and that on 27 April 2026 a magistrate judge granted a stipulated order under which the state agreed not to enforce SB 24-205 pending a ruling on a preliminary-injunction motion (McDermott Will & Emery, 2026) [VERIFICATION NEEDED: litigation details from the law-firm alert only; the court docket was not checked; the final status of the case is unknown].

## 6. Applications & Concrete Examples
Concrete effects: federal procurement of AI follows executive direction; state laws such as California's SB 53 impose duties on frontier developers; the Department of Justice task force is a mechanism for litigation against state laws. The Colorado litigation described in Section 5 involved a Department of Justice filing, but no ruling on the merits of any challenge to a state AI law was found in the sources reviewed, and none is recorded.

## 7. Advantages, Limitations & Unresolved Questions
- *Argued advantages*: speed and flexibility, a single national standard, lower compliance burden.
- *Limitations*: executive measures are reversible and lack statutory force; litigation over preemption raises constitutional questions (legal commentators expect prolonged uncertainty per secondary reports); absence of binding federal safety requirements for frontier models; reliance on voluntary industry commitments.
- *Contested*: whether federal preemption without a substantive federal standard is desirable or lawful; whether training on copyrighted material is lawful (the 2026 framework states an administration view but leaves fair use to courts).

## 8. Current State of Knowledge
As of 2026-10-18: Executive Order 14110 is revoked; Order 14179 and the July 2025 Action Plan state current federal direction; Order 14365 (December 2025) is an executive instrument (its implementation status was not checked); the March 2026 framework is a recommendation only; no comprehensive federal AI statute or preemption statute has been located. State laws, including California's SB 53, are not known to have been preempted or enjoined in the sources reviewed; the Colorado AI Act was replaced by SB 26-189 in May 2026.

## 9. Future Trajectory & Strategic Outlook
Observed: policy moves by executive action and litigation rather than statute. No forecasts are registered here. Items to watch: congressional preemption bills, court rulings on Order 14365 and on Colorado's law, and revisions to export-control rules. See Living Reference `TIME-001`.

## 10. Related Entries & Knowledge Graph
- **Foundational Concepts**:
  - [[GOV-LAW-001] Global AI Governance & Regulation](../vol08_safety_governance/GOV-LAW-001_ai_governance.md): general context; note that GOV-LAW-001 describes the revoked Order 14110 as operative and should be read with this article.
- **Comparative regimes**:
  - [[GOV-EU-001] The EU AI Act](../vol08_safety_governance/GOV-EU-001_the_eu_ai_act.md)
  - [[GOV-CN-001] China's AI Regulation](../vol08_safety_governance/GOV-CN-001_chinas_ai_regulation.md)
- **Related**:
  - [[GOV-EXPORT-001] Export Controls and AI Geopolitics](../vol08_safety_governance/GOV-EXPORT-001_export_controls_and_ai_geopolitics.md): chip and model controls.
  - [[GOV-INTL-001] International AI Governance and Compute Governance](../vol08_safety_governance/GOV-INTL-001_international_ai_governance_and_compute_governance.md)
  - [[ORG-SAFETY-001] AI Safety Institutes and Research Organizations](../vol10_people_organizations/ORG-SAFETY-001_ai_safety_institutes_and_research_organizations.md)
  - [[ORG-STANDARDS-001] Standards Bodies (ISO/IEC, NIST)](../vol10_people_organizations/ORG-STANDARDS-001_standards_bodies_iso_iec_nist.md)

## 11. References & Foundational Literature
1. Federal Register. (2025a). *Initial Rescissions of Harmful Executive Orders and Actions* (Executive Order 14148, signed 20 January 2025). 90 FR 8237 (28 January 2025), FR Doc. 2025-01901. https://www.federalregister.gov/documents/2025/01/28/2025-01901/initial-rescissions-of-harmful-executive-orders-and-actions. [Sec. 2(ggg) revokes Executive Order 14110; read from the GovInfo PDF]. Also published at https://www.whitehouse.gov/presidential-actions/2025/01/initial-rescissions-of-harmful-executive-orders-and-actions/.
2. Federal Register. (2025b). *Removing Barriers to American Leadership in Artificial Intelligence* (Executive Order 14179, signed 23 January 2025). 90 FR 8741 (31 January 2025), FR Doc. 2025-02172. https://www.federalregister.gov/documents/2025/01/31/2025-02172/removing-barriers-to-american-leadership-in-artificial-intelligence. [Sec. 2 policy statement, Sec. 4 180-day action plan, Sec. 5 review of Order 14110 actions].
3. Federal Register. (2025c). *Ensuring a National Policy Framework for Artificial Intelligence* (Executive Order 14365, signed 11 December 2025). 90 FR 58499 (16 December 2025), FR Doc. 2025-23092. https://www.federalregister.gov/documents/2025/12/16/2025-23092/ensuring-a-national-policy-framework-for-artificial-intelligence. [Litigation task force, State-law evaluation, BEAD conditions, FCC and FTC directives, legislative recommendation].
4. The White House. (2025d). *Winning the Race: America's AI Action Plan* (July 2025). https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf. [Three-pillar structure and pillar titles in the table of contents].
5. The White House. (2025e). *White House Unveils America's AI Action Plan* (23 July 2025). https://www.whitehouse.gov/articles/2025/07/white-house-unveils-americas-ai-action-plan/. [Publication date of the announcement].
6. The White House. (2025f). *Fact Sheet: President Donald J. Trump Ensures a National Policy Framework for Artificial Intelligence* (11 December 2025). https://www.whitehouse.gov/fact-sheets/2025/12/fact-sheet-president-donald-j-trump-ensures-a-national-policy-framework-for-artificial-intelligence/. [Summary of Order 14365; the order text in Ref 3 is the primary source used here; fact-sheet page not re-fetched in the 2026-10-18 pass].
7. Federal Register. (2023). *Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence* (Executive Order 14110, signed 30 October 2023). 88 FR 75191 (1 November 2023), FR Doc. 2023-24283. https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence. [Sec. 4.2 reporting requirements and interim 10^26 / 10^23 thresholds; read from the GovInfo PDF].
8. National Institute of Standards and Technology. (2023). *Artificial Intelligence Risk Management Framework (AI RMF 1.0)*, NIST AI 100-1, January 2023. https://doi.org/10.6028/NIST.AI.100-1. [Voluntary US risk-management framework; DOI resolves to the NIST PDF dated January 2023].
9. National Institute of Standards and Technology. (n.d.). *Center for Advancing Innovation and Standards for Super Intelligence (CAISSI)*. https://www.nist.gov/caissi (formerly /caisi), read 2026-10-18, last modified 3 September 2026. [Current NIST name and mission statement of the body].
10. U.S. Government Publishing Office. (2021). *Public Law 116-283, William M. (Mac) Thornberry National Defense Authorization Act for Fiscal Year 2021* (1 January 2021), Division E, National Artificial Intelligence Initiative Act of 2020. https://www.govinfo.gov/content/pkg/PLAW-116publ283/html/PLAW-116publ283.htm. [Division E title and law date; table of contents only was read].
11. Office of the Governor of California. (2025). *Governor Newsom signs SB 53, advancing California's world-leading artificial intelligence industry* (29 September 2025). https://www.gov.ca.gov/2025/09/29/governor-newsom-signs-sb-53-advancing-californias-world-leading-artificial-intelligence-industry/. [Signing date and summary of requirements].
12. Colorado General Assembly. (2026). *SB26-189, Automated Decision-Making Technology*. https://leg.colorado.gov/bills/sb26-189. [Signed 14 May 2026; repeals and re-enacts SB 24-205 provisions].
13. McDermott Will & Emery. (2026). *Colorado AI law in flux: comprehensive replacement bill signed after federal court blocks predecessor's enforcement* (27 May 2026). https://www.mcdermottlaw.com/insights/colorado-ai-law-in-flux-comprehensive-replacement-bill-signed-after-federal-court-blocks-predecessors-enforcement/. [Tier 4 legal commentary: xAI suit, DOJ complaint, stipulated order].
14. WilmerHale. (2026). *White House Releases National Policy Framework for Artificial Intelligence* (23 March 2026). https://www.wilmerhale.com/en/insights/blogs/wilmerhale-privacy-and-cybersecurity-law/20260323-white-house-releases-national-policy-framework-for-artificial-intelligence. [Tier 4 legal commentary: 20 March 2026 framework, non-binding, preemption scope].
15. StateScoop. (2025a). *State AI moratorium dies in Senate* (1 July 2025). https://statescoop.com/state-ai-moratorium-dies-senate/. [Tier 4 reporting: 99-1 Senate vote to remove the moratorium; Senate roll call itself not retrieved (senate.gov returned 403)].
16. StateScoop. (2025b). *State AI law moratorium omitted from 2026 defense bill, but Trump is preparing 'ONE RULE' executive order* (8 December 2025). https://statescoop.com/state-ai-law-moratorium-omitted-2026-defense-bill-trump-eo/. [Tier 4 reporting: NDAA text did not include the moratorium].
17. National Institute of Standards and Technology. (2026, October 16). *Key Practices for Incident Readiness and Response in Super Intelligence Training and Evaluation*. https://www.nist.gov/news-events/news/2026/10/key-practices-incident-readiness-and-response-super-intelligence-training, read 2026-10-18. [NIST news item that refers to "NIST's Center for Advancing Innovation and Standards for Super Intelligence (NIST CAISSI)"].

## 12. Evidential Classification Matrix
| Section / Claim | Evidence Category | Supporting Source | Confidence / Status |
| :--- | :--- | :--- | :--- |
| EO 14110 revoked (EO 14148, 20 Jan 2025) | Established | Ref 1, Ref 2 | High |
| EO 14110 Sec. 4.2 thresholds | Established (legal record) | Ref 7 | High |
| EO 14179 content | Established | Ref 2 | High |
| EO 14365 directives | Established | Ref 3 | High for text; implementation status unchecked |
| Action Plan pillars and date | Well Supported | Ref 4, Ref 5 | High for structure; content beyond table of contents not read |
| CAISI rename (June 2025) | Preliminary | trade press; Ref 9 for current name | Moderate; primary release not retrieved |
| March 2026 framework non-binding | Preliminary | Ref 14 | Moderate; primary text not read |
| No federal preemption statute | Preliminary | Ref 15, Ref 16 | Moderate; congress.gov not accessible |
| SB 53 signed 29 Sep 2025 | Established | Ref 11 | High; threshold and effective date from law firms only (Moderate) |
| Colorado SB 26-189 | Established | Ref 12 | High for signing; litigation details Moderate (Ref 13) |
| Outcome of preemption litigation | Contested / Unknown | none | Unresolved |

## 13. Revision History
- **v0.1.0** (2026-10-15): Initial draft. Citations checked against official pages where marked; not independently fact-checked.
- **v0.2.0** (2026-10-18): Fact-check pass (see `docs/audits/GOV-US-001_audit.md`). EO numbers, dates and Federal Register citations verified against the Federal Register/GovInfo; EO 14110 thresholds and EO 14365 directives restated from the order text; CAISI/CAISSI wording corrected; Colorado SB 26-189 and SB 53 signing added; NIST news item of 16 October 2026 retained (page read directly); references renumbered.