Executive Order on AI
Last reviewed
Sources
48 citations
Review status
Source-backed
Revision
v5 ยท 8,461 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
48 citations
Review status
Source-backed
Revision
v5 ยท 8,461 words
Add missing citations, update stale details, or suggest a clearer explanation.
The Executive Order on AI most commonly refers to Executive Order 14110, titled "Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence," signed by US President Joe Biden on October 30, 2023. It was the most comprehensive piece of artificial intelligence governance issued by the United States government, directing over 50 federal entities to undertake more than 100 specific actions across eight policy areas related to AI safety, security, equity, innovation, and international cooperation. The order was revoked by President Donald Trump on January 20, 2025, and replaced three days later by Executive Order 14179, "Removing Barriers to American Leadership in Artificial Intelligence," which shifted the federal government's AI posture from risk mitigation and oversight toward deregulation and innovation promotion [1][2].
Since January 2025, the Trump administration has issued a sequence of additional executive orders, agency directives, and a 90-plus action policy roadmap (America's AI Action Plan, July 2025) that together form a new federal framework: pro-innovation, pro-export, infrastructure-heavy, and openly skeptical of what it labels "woke" content moderation in AI. The most recent action, a June 2, 2026 executive order titled "Promoting Advanced Artificial Intelligence Innovation and Security," asks frontier AI developers to give the federal government access to leading-edge models up to 30 days before public release for a national security and cybersecurity review. This article covers the Biden-era order, the post-2025 Trump framework, and the key personnel, EOs, and agency actions that emerged in 2025 and 2026 [2][45].
Executive Order 14110 was the United States' first attempt at a comprehensive, government-wide AI policy framework. Rather than a single rule, it functioned as an orchestration document: it set eight guiding principles, then assigned more than 100 deliverables (drafting standards, issuing guidance, reporting to Congress, building institutions) to over 50 federal entities on staggered deadlines. Its most legally novel feature was using the Defense Production Act to compel private AI developers to report safety-test results and computing-cluster information to the government. The Trump-era EO 14179 framework keeps the same orchestration model but inverts the goals: instead of compelled safety reporting, it directs agencies to repeal regulations, accelerate energy and data center permitting, promote AI exports, and procure only "ideologically neutral" models.
By mid-2023, the rapid proliferation of large language models such as GPT-4, Claude, and Gemini had created intense pressure on governments worldwide to respond to both the opportunities and risks posed by advanced AI. In the United States, there was no comprehensive federal AI legislation. The Biden administration had taken incremental steps, including the Blueprint for an AI Bill of Rights (October 2022) and voluntary commitments from 15 leading AI companies in July 2023 to manage AI risks. However, these measures were widely seen as insufficient given the pace of technological change [3].
Internationally, the European Union was advancing the EU AI Act, which would become the world's first comprehensive AI law upon its formal adoption in 2024. China had already implemented binding regulations on generative AI services, effective August 2023. Against this backdrop, the Biden administration sought to demonstrate American leadership on AI governance without waiting for Congress to pass legislation [4].
The executive order leveraged presidential authority, including powers under the Defense Production Act (DPA), to impose requirements on AI developers and direct federal agencies to take action. The use of the DPA was significant because it gave the order legal teeth beyond what a typical executive directive could achieve: under DPA Title VII, the government could compel private companies to provide information about their AI development activities [5].
The order was organized around eight guiding principles and covered an extraordinarily broad range of AI-related policy areas. Its most consequential provisions fell into several categories.
The order defined "dual-use foundation models" as AI models that exhibit, or could be easily modified to exhibit, high levels of performance at tasks posing serious risks to security, public health, or safety. It required developers of such models to report to the federal government on an ongoing basis, sharing information about the development process, cybersecurity measures, ownership of model weights, and results of safety testing [1].
The reporting threshold was set at models trained using more than 10^26 integer or floating-point operations (FLOPs). A Biden administration official stated at the time that the threshold was calibrated so that "current models wouldn't be captured but the next generation state-of-the-art models likely would." For context, GPT-3's 175 billion parameter version required roughly 318 times less computing power than this threshold. The threshold was intended to be updated by the Secretary of Commerce as the state of the art advanced [5].
The order also required companies possessing large-scale computing clusters (defined as having a theoretical maximum computing capacity of 10^20 FLOPs per second for AI training) to report the location and total computing power of those clusters to the federal government [5].
Developers of the most powerful AI systems were required to share the results of safety testing and red-teaming exercises with the federal government. Red-teaming, the practice of deliberately testing AI systems for vulnerabilities and harmful capabilities, was to be conducted according to standards developed by NIST (the National Institute of Standards and Technology). This provision aimed to ensure that the government had visibility into the safety profiles of frontier models before they were widely deployed [1].
The order directed the creation of the US AI Safety Institute within NIST, tasked with developing standards, guidelines, and best practices for AI safety and security. The institute was charged with several responsibilities:
| Responsibility | Description |
|---|---|
| Safety standards | Develop guidelines for safe, secure, and trustworthy AI systems |
| Evaluation frameworks | Create benchmarks and testing methodologies for assessing AI model capabilities and risks |
| Red-team guidance | Establish standards for red-teaming and adversarial testing of AI systems |
| Content authentication | Develop guidance on watermarking and content provenance for AI-generated content |
| International cooperation | Collaborate with international counterparts, including the UK AI Safety Institute |
The AI Safety Institute quickly became one of the most prominent outcomes of the executive order, building evaluation teams and conducting joint assessments with its UK counterpart. In August 2024, the institute signed memoranda of understanding with Anthropic and OpenAI for pre-deployment access to frontier models. These agreements survived the change in administration in modified form (see below) [6].
The order directed the Department of Commerce to develop guidance for content authentication and watermarking to clearly label AI-generated content. Watermarking was defined in the order as "the act of embedding information, which is typically difficult to remove, into outputs created by AI, including into outputs such as photos, videos, audio clips, or text, for the purposes of verifying the authenticity of the output or the identity or characteristics of its provenance, modifications, or conveyance" [1].
Recognizing that the United States depended heavily on foreign-born researchers and engineers for its AI workforce, the order directed the State Department and Department of Homeland Security to streamline visa processes for AI experts. Specific measures included modernizing visa criteria for highly skilled AI professionals, clarifying pathways for researchers and entrepreneurs, and reducing processing times. The order also directed agencies to consider how immigration policy could help the US attract and retain global AI talent [1].
The order required each federal agency to designate a Chief Artificial Intelligence Officer (CAIO) responsible for overseeing the agency's use of AI. Agencies were directed to take a risk-based approach to deploying generative AI rather than imposing blanket bans. They were also required to complete inventories of their AI use cases and ensure that AI systems used in government did not discriminate against protected groups. These obligations were operationalized through OMB Memorandum M-24-10 (March 2024) on advancing governance, innovation, and risk management for federal AI use, and M-24-18 (September 2024) on federal AI procurement [1].
The order also addressed a wide range of other topics:
| Area | Key Requirements |
|---|---|
| Civil rights | Guidance to prevent algorithmic discrimination in hiring, lending, criminal justice, and other domains |
| Workers | Studies on AI's impact on the labor market; guidance on employer use of AI for surveillance and evaluation |
| Healthcare | Development of AI safety standards for healthcare applications |
| Education | Recommendations for using AI in educational settings |
| Privacy | Research into privacy-preserving techniques; assessment of how AI changes the privacy landscape |
| Competition | Actions to promote competition in AI markets and prevent monopolistic concentration |
| National security | Classified companion memo on AI and national security |
Between its signing in October 2023 and its revocation in January 2025, Executive Order 14110 produced substantial implementation activity. NIST published its AI Risk Management Framework Generative AI Profile (NIST AI 600-1) in July 2024. The Department of Commerce established initial reporting requirements for computing clusters. Multiple agencies appointed Chief AI Officers. The AI Safety Institute began operations, hiring staff and conducting evaluations of frontier models, including a joint evaluation of OpenAI's o1 model with the UK AI Safety Institute [6][7].
However, many of the order's mandates had aggressive timelines (90 days, 180 days, 270 days) that proved challenging for federal agencies to meet. Some deadlines were missed, and the breadth of the order meant that implementation was uneven across agencies [7].
On January 20, 2025, within hours of taking office for his second term, President Donald Trump revoked Executive Order 14110 as part of a broader package of executive actions reversing Biden-era policies. Three days later, on January 23, 2025, Trump signed Executive Order 14179, titled "Removing Barriers to American Leadership in Artificial Intelligence." The order was published in the Federal Register on January 31, 2025 [2][11].
The Trump order stated that it aimed to "strengthen United States leadership in artificial intelligence," promote AI development "free from ideological bias or social agendas," and establish an action plan to maintain global AI dominance. The order reflected the incoming administration's view that Biden's AI governance approach was overly regulatory and risked hampering American competitiveness, particularly relative to China [2].
EO 14179 is short, six sections in total, and primarily directive rather than prescriptive. The substantive content sits in three sections [2][11]:
| Section | Content |
|---|---|
| Section 2 (Policy) | Declares the policy of the United States is to "sustain and enhance America's global AI dominance" for human flourishing, economic competitiveness, and national security |
| Section 3 (Definition) | Adopts the AI definition from 15 U.S.C. 9401(3) (the National AI Initiative Act of 2020) |
| Section 4 (AI Action Plan) | Directs the Assistant to the President for Science and Technology, the Special Advisor for AI and Crypto, and the Assistant to the President for National Security Affairs to develop an AI Action Plan within 180 days |
| Section 5 (Implementation) | Orders an immediate review of all policies and actions taken under EO 14110, requires identification of those that are "inconsistent with, or present obstacles to" the new policy, and directs the Office of Management and Budget to revise OMB Memoranda M-24-10 and M-24-18 within 60 days |
The order does not name individuals, but the three coordinators identified in Section 4 were filled by Michael Kratsios (Office of Science and Technology Policy director, confirmed March 25, 2025), David Sacks (Special Advisor for AI and Crypto), and the National Security Advisor (Mike Waltz initially, then Marco Rubio in an acting capacity from May 2025) [12][13].
The revocation eliminated several key elements of the Biden order:
| Revoked Element | Consequence |
|---|---|
| Dual-use foundation model reporting | Companies no longer required to share safety testing results or development information with the government |
| Computing cluster reporting | Reporting requirements for large-scale computing infrastructure eliminated |
| Red-teaming mandates | Government-required red-teaming standards no longer in effect |
| Chief AI Officer requirement | Federal agencies no longer mandated to appoint CAIOs (though some retained them voluntarily) |
| Watermarking guidance | Department of Commerce efforts on content authentication were deprioritized |
| Immigration streamlining | AI talent visa initiatives were rolled back as part of the broader immigration policy shift |
Not everything from the Biden era was eliminated. The NIST AI Risk Management Framework, published in January 2023 (before Executive Order 14110), remained in place as a voluntary, non-binding guidance document. Many private companies continued to use it as a best-practice reference. Some agencies retained their Chief AI Officers and AI governance structures even without the federal mandate. And corporate safety commitments made at the AI Safety Summit series (the Seoul Frontier AI Safety Commitments, for instance) existed independently of any US executive order [8].
However, the US AI Safety Institute at NIST underwent significant changes. In June 2025, Secretary of Commerce Howard Lutnick announced that the institute would be renamed the Center for AI Standards and Innovation (CAISI), dropping "safety" from its title and reorienting its mission toward national security risks and global competitiveness. Lutnick stated: "For far too long, censorship and regulations have been used under the guise of national security. Innovators will no longer be limited by these standards" [9][14].
The Trump AI policy team is concentrated in three roles, all of whom are credited as authors of the July 2025 AI Action Plan [12][13][15].
| Position | Name | Notes |
|---|---|---|
| Director, Office of Science and Technology Policy (OSTP) | Michael Kratsios | First-term Trump CTO, confirmed by Senate 74 to 25 on March 25, 2025; also serves as Assistant to the President for Science and Technology |
| Special Advisor for AI and Crypto ("AI Czar") | David Sacks | PayPal founding-era COO, venture capitalist, named December 5, 2024; role limited to a 130-day Special Government Employee window. Stepped down in March 2026 to co-chair the President's Council of Advisors on Science and Technology (PCAST) [16] |
| Senior White House Policy Advisor for Artificial Intelligence | Sriram Krishnan | Former Andreessen Horowitz partner, named December 22, 2024, full-time White House staff role |
Vice President JD Vance has also been a public face of the administration's AI policy, most prominently at the Paris AI Action Summit in February 2025. The administration has not appointed a single "AI Secretary" or stand-alone AI office; instead, AI policy is coordinated across OSTP, the National Security Council, OMB, and the Department of Commerce.
One day after Trump's inauguration and two days before signing EO 14179, the President announced the Stargate Project at a White House press conference alongside OpenAI CEO Sam Altman, Oracle chairman Larry Ellison, and SoftBank CEO Masayoshi Son. The joint venture, formally Stargate LLC, planned to invest up to 500 billion USD in US AI infrastructure by 2029, with an initial 100 billion USD commitment. Initial reporting suggested SoftBank and OpenAI would each contribute 19 billion USD for 40 percent ownership, with Oracle and Abu Dhabi's MGX taking smaller stakes. Stargate was framed as a private-sector counterpart to the executive orders: the government would streamline permitting and energy approvals, and the private sector would build [17].
On February 11, 2025, Vice President JD Vance delivered the United States' keynote remarks at the Paris AI Action Summit hosted by French President Emmanuel Macron. Vance opened by stating that he was "not here this morning to talk about AI safety, which was the title of the conference a couple of years ago," and was instead there "to talk about AI opportunity." He laid out four priorities for US policy [18][19]:
The speech was notable for several things it did not do. The United States and the United Kingdom both declined to sign the summit's joint statement on "inclusive and sustainable" AI, which 60 other countries (including France, Germany, Canada, India, and China) did sign. Vance criticized European AI regulation by name, warning that "tightening the screws on US tech companies with international footprints" would "chill the development of AI worldwide." The speech is widely seen as the moment the international AI safety summit series, launched by the UK at Bletchley Park in November 2023, lost US support as a coordinating venue [18][19].
The AI Action Plan required by EO 14179, formally titled "Winning the Race: America's AI Action Plan," was released on July 23, 2025, on the 181st day after the executive order, just inside the 180-day deadline. The plan contains over 90 federal policy actions organized into three pillars [20][21][22].
The first pillar emphasizes deregulation and removing barriers to AI development:
The second pillar focuses on physical infrastructure, primarily energy, data centers, and semiconductors:
The third pillar mobilizes the Departments of Commerce and State on AI exports and standards:
| Agency | Selected deliverables |
|---|---|
| Office of Science and Technology Policy (OSTP) | Coordination of plan implementation; Director Kratsios as administration lead spokesperson |
| Department of Commerce / NIST / CAISI | Revise the AI Risk Management Framework; lead AI export package program; develop AI Agent Standards Initiative (announced February 2026); test frontier models pre-deployment |
| Department of Energy (DOE) | Identify federal sites for data centers; release AI Strategy and Compliance Plan (issued October 2, 2025); accelerate nuclear and geothermal deployment for AI loads |
| Department of Defense (DOD) | Refine Responsible AI and Generative AI frameworks; build high-security data centers for intelligence use; collaborate with NIST and ODNI |
| Department of Homeland Security (DHS) | Lead the AI Information Sharing and Analysis Center; issue private-sector incident response guidance |
| Department of the Interior (DOI) | Authorize data center construction on federal lands |
| National Science Foundation (NSF) | Expand compute access for academic researchers and small businesses; AI education research |
| Department of Education / Department of Labor | Implement AI literacy curriculum and apprenticeship pathways under EO 14277 (see below) |
| Office of Management and Budget (OMB) | Issue federal procurement guidance within 120 days; revise M-24-10 and M-24-18 to align with EO 14179 |
| Federal Trade Commission | Issue policy statement on FTC Act preemption of conflicting state AI laws (per EO of December 11, 2025) |
EO 14179 is the umbrella, but the operative legal pressure comes from a sequence of follow-on executive orders. The most consequential are summarized below.
Signed the same day as the revocation of EO 14110, this order declared a national energy emergency under the National Emergencies Act, citing the need to meet projected demand from "manufacturing, re-industrialization, and data centers driving artificial intelligence (AI) innovation." It directs federal agencies to invoke emergency authorities to fast-track energy infrastructure, including on federal lands, and to consider use of the Defense Production Act and federal eminent domain. The emergency was renewed for another year on January 14, 2026 [23].
In May 2025, the Department of Commerce rescinded the Biden administration's January 2025 AI diffusion rule, a tiered chip export framework that had organized countries into three categories with differing license requirements for advanced AI chips from companies including Nvidia, AMD, and Intel. The rule never took effect [24].
This EO, sometimes referred to as EO 14277, establishes the White House Task Force on Artificial Intelligence Education chaired by the OSTP Director and including the Secretaries of Agriculture, Labor, Energy, and Education, the NSF Director, the Special Advisor for AI and Crypto, and others. It directs the Department of Education to prioritize funding for teacher AI training, the Department of Labor to expand AI-related apprenticeships, and the NSF to escalate research on AI in education. It also calls for boosting AI courses and certificate programs available to high school students [25].
On the day the AI Action Plan was released, Trump signed three accompanying executive orders [26][27]:
| EO | Title | Key provisions |
|---|---|---|
| 14318 | Accelerating Federal Permitting of Data Center Infrastructure | Defines a "Data Center Project" as a facility requiring more than 100 megawatts of new load dedicated to AI; creates new NEPA categorical exclusions; directs Interior, Energy, and Defense to authorize construction on federal lands. The DOE selected four initial sites the next day: Idaho National Laboratory, Oak Ridge Reservation (Tennessee), Paducah Gaseous Diffusion Plant (Kentucky), and the Savannah River Site (South Carolina) |
| 14319 | Preventing Woke AI in the Federal Government | Establishes two "Unbiased AI Principles" for federal LLM procurement: truth-seeking (prioritizing "historical accuracy, scientific inquiry, and objectivity") and ideological neutrality (LLMs as "neutral, nonpartisan tools that do not manipulate responses in favor of ideological dogmas such as DEI"). OMB must issue implementation guidance within 120 days. Includes carve-outs for national security systems |
| 14320 | Promoting the Export of the American AI Technology Stack | Establishes the American AI Exports Program; directs the Secretary of Commerce to evaluate full-stack AI export proposals by October 21, 2025; designates selected packages as "priority AI export packages" eligible for federal financing through the Export-Import Bank, the Development Finance Corporation, and similar agencies |
In October 2025, the US Air Force followed up on EO 14318 by soliciting proposals to develop "underutilized" lands at five bases (Arnold AFB in Tennessee, Edwards AFB in California, Joint Base McGuire-Dix-Lakehurst in New Jersey, Davis-Monthan AFB in Arizona, and Robins AFB in Georgia) for data center use [28].
On December 11, 2025, Trump signed an executive order, formally titled "Ensuring a National Policy Framework for Artificial Intelligence" (also referred to in some White House materials as "Eliminating State Law Obstruction of National Artificial Intelligence Policy"), that targets state-level AI regulation. Its principal mechanisms are [29][30][31]:
The order carves out state laws on child safety, AI compute and data center infrastructure, and state procurement and governmental use of AI from proposed preemption [29][30].
On June 2, 2026, Trump signed an executive order titled "Promoting Advanced Artificial Intelligence Innovation and Security," the administration's first to focus squarely on frontier model evaluation and AI-enabled cyber threats. Its central mechanism is a voluntary 30-day pre-release review: AI developers are asked to provide the federal government access to "covered frontier models" for a national security and cybersecurity assessment up to 30 days before the models are released to other trusted partners. A classified benchmark, rather than a public compute threshold, determines which models are "covered" [45][46].
The order grew directly out of a concrete capability demonstration. On May 22, 2026, Anthropic published its first Project Glasswing update, reporting that a preview of its Claude Mythos model had autonomously identified more than 10,000 high- or critical-severity software vulnerabilities in a single month, including a critical wolfSSL flaw (CVE-2026-5194) affecting roughly 5 billion devices. The order establishes a Treasury-led clearinghouse to coordinate vulnerability discovery and patch dissemination across AI firms, technology companies, and critical infrastructure operators, and routes frontier-model evaluation toward national security agencies including the Treasury Department, the National Security Agency, and the Cybersecurity and Infrastructure Security Agency (CISA) [45][46].
A significant side effect is that CAISI was directed to stop publishing its model-evaluation findings, ending the public track record it had built since the AI Safety Institute era. Critics argued the shift trades transparency for secrecy; the administration framed it as protecting national-security-sensitive capability data. The order keeps the broader deregulatory posture of EO 14179 intact while adding a security overlay focused on AI's offensive cyber potential [45][47].
| Date | Action | Effect |
|---|---|---|
| October 30, 2023 | EO 14110 signed by Biden | Comprehensive Biden AI executive order |
| December 5, 2024 | David Sacks named AI and Crypto Czar | Signaled the incoming administration's AI policy direction |
| December 22, 2024 | Sriram Krishnan named Senior White House AI Policy Advisor | Filled out the senior AI team |
| January 20, 2025 | EO 14110 revoked; EO 14156 (Energy Emergency) signed | Cleared the regulatory deck; emergency posture for energy |
| January 21, 2025 | Stargate Project announced (500 billion USD) | Private-sector counterpart to deregulation |
| January 23, 2025 | EO 14179 signed | New umbrella AI executive order; 180-day deadline for action plan |
| February 11, 2025 | JD Vance speech at Paris AI Action Summit | US declines to sign summit communique |
| March 25, 2025 | Michael Kratsios confirmed as OSTP Director | Senate vote 74-25 |
| April 23, 2025 | EO 14277 (AI Education for American Youth) | White House Task Force on AI Education |
| May 2025 | AI diffusion chip rule rescinded | Reversal of Biden export control framework |
| June 2025 | US AI Safety Institute renamed CAISI | Mission shift toward standards and competitiveness |
| July 4, 2025 | OBBBA signed without state AI moratorium | Senate stripped 10-year preemption 99 to 1 |
| July 23, 2025 | AI Action Plan released; EOs 14318, 14319, 14320 signed | Permitting, anti-"woke" procurement, AI exports |
| October 2, 2025 | DOE AI Strategy and Compliance Plan released | Agency-level implementation of the plan |
| December 11, 2025 | EO on National Policy Framework for AI | State preemption push, DOJ litigation task force |
| January 10, 2026 | DOJ AI Litigation Task Force operational | Begins challenging state AI laws in federal court |
| February 2026 | NIST AI Agent Standards Initiative announced | Standards work continues at CAISI |
| May 5, 2026 | CAISI testing agreements with Google DeepMind, Microsoft, xAI | Joins existing renegotiated deals with Anthropic and OpenAI |
| June 2, 2026 | EO "Promoting Advanced AI Innovation and Security" | Voluntary 30-day frontier model review; CAISI evaluations move to classified framework |
| Dimension | Biden EO 14110 (Oct 2023, revoked) | Trump EO 14179 framework (Jan 2025 onward) |
|---|---|---|
| Core posture | Risk mitigation, oversight, equity | Innovation, dominance, deregulation |
| Authority used | Defense Production Act for compelled reporting | Procurement, permitting, federal lands, export controls |
| Industry reporting | Mandatory for dual-use foundation models above 10^26 FLOPs | Voluntary CAISI testing agreements; voluntary 30-day pre-release review (June 2026) |
| Federal procurement | Risk-based; equity reviews | Two "unbiased AI principles": truth-seeking and ideological neutrality |
| Key institutions | US AI Safety Institute, CAIOs, AI Bill of Rights | CAISI, Chief AI Officer Council, AI Litigation Task Force |
| State law posture | Permissive; states free to regulate | Preemption push via DOJ, FTC, BEAD funding leverage |
| Energy and infrastructure | Limited focus | Central focus: emergency declaration, NEPA exclusions, federal land use |
| Immigration | Streamline AI talent visas | Rolled back |
| International coordination | Bletchley/Seoul summit series; UK AISI partnership | Paris summit communique declined; bilateral CAISI-AISI cooperation continues |
The contrast between the Biden executive order's approach and the EU AI Act highlights fundamental differences in how the two jurisdictions approached AI governance.
| Dimension | Biden Executive Order 14110 | EU AI Act |
|---|---|---|
| Legal authority | Executive action (revocable by next president) | Binding legislation passed by European Parliament and Council |
| Durability | Revoked after 15 months | Permanent law with phased implementation through 2026 |
| Enforcement mechanism | DPA reporting requirements; agency directives | Fines up to 35 million EUR or 7% of global revenue |
| Risk classification | Focused on dual-use foundation models above 10^26 FLOPs | Four-tier risk system: unacceptable, high, limited, minimal |
| Scope | Broad (safety, equity, labor, privacy, immigration, national security) | Focused on AI system lifecycle (development, deployment, use) |
| Transparency requirements | Developer reporting to government | Mandatory transparency for users interacting with AI; public documentation |
| Approach to innovation | Balanced safety with innovation promotion | Risk-based regulation with exemptions for research and open-source |
| International influence | Influential but short-lived | Widely seen as a regulatory model for other jurisdictions |
The EU AI Act's risk-based classification system categorizes AI applications into four tiers, with different regulatory requirements for each. Systems posing "unacceptable risk" (such as social scoring and certain biometric surveillance applications) are prohibited outright. High-risk systems (including AI used in critical infrastructure, education, employment, and law enforcement) face stringent requirements including conformity assessments, documentation, and human oversight. General-purpose AI models, including large language models, face additional transparency and safety requirements, particularly those exceeding 10^25 FLOPs in training compute [4].
The Biden executive order, by contrast, was primarily focused on the most capable frontier models and did not establish a comprehensive classification system for all AI applications. Its strength lay in its breadth (covering immigration, labor, education, and national security alongside pure safety concerns) and in the reporting requirements enabled by the Defense Production Act. Its weakness was its impermanence: as an executive order, it could be (and was) revoked by the next president [4].
Vance's Paris speech and the subsequent AI Action Plan widened the gap between US and EU approaches. By late 2025, the European Commission was reportedly considering softening parts of the AI Act amid industry lobbying and Trump administration pressure. In November 2025, lawmakers including MEP Brando Benifei warned the Commission against "watering down" the AI Act "to appease Trump and US tech companies." In December 2025, the Trump administration imposed visa restrictions on European officials it accused of pressuring US tech firms to censor American speech, prompting French President Emmanuel Macron to call the measures "intimidation and coercion aimed at undermining European digital sovereignty" [32][33].
In June 2025, Secretary of Commerce Howard Lutnick announced that the US AI Safety Institute would be renamed the Center for AI Standards and Innovation (CAISI) and reoriented toward national security and US competitiveness rather than broad AI safety. CAISI remained inside NIST and inside the Commerce Department's organizational chart, and retained much of the institute's evaluation and standards-setting capacity, but staff layoffs reduced research personnel hired under the Biden mandate [9][14][34].
CAISI's testing relationships with frontier labs continued in modified form. The Anthropic and OpenAI memoranda of understanding, originally signed with the AISI in August 2024, were renegotiated in 2025 to align with AI Action Plan priorities. On May 5, 2026, CAISI signed parallel testing agreements with Google DeepMind, Microsoft, and xAI, expanding its frontier evaluation program to five major labs. Under the agreements, the labs share pre-deployment access to frontier models, including versions with safety guardrails reduced or removed, for national security testing in classified environments. CAISI publicly reported having completed more than 40 model assessments by mid-2026, covering cybersecurity, biosecurity, and chemical weapons risks; in April 2026 it evaluated the open-weight DeepSeek V4 Pro and concluded the model lagged the frontier by roughly eight months [34][48].
In February 2026, NIST announced the AI Agent Standards Initiative for interoperable and secure AI agent development, a workstream that emerged directly from the AI Action Plan [22]. Following the June 2, 2026 executive order, CAISI was directed to stop publishing its evaluation findings, with frontier model review routed to a classified process led by national security agencies [45][47].
The US rebrand was paralleled in the United Kingdom. On February 14, 2025, three days after Vance's Paris speech, UK Technology Secretary Peter Kyle announced at the Munich Security Conference that the UK AI Safety Institute would be renamed the AI Security Institute and refocused on cyber risks rather than AI ethics or content moderation. The two institutes signed an interoperability agreement in April 2025, and joint testing exercises and shared taxonomies of AI risk continued through 2026 [35].
Lina Khan's tenure as FTC chair ended on January 20, 2025. Trump elevated sitting Republican commissioner Andrew Ferguson to chair, a move that did not require Senate confirmation. Ferguson stated that the FTC would stop "legally dubious" consumer protection cases and end efforts to become an "AI regulator," while continuing antitrust scrutiny of AI mergers. In October 2025, the FTC quietly removed several Khan-era blog posts on AI risks and open-weight models. Under the December 11, 2025 executive order, Ferguson was directed to issue a policy statement, by March 11, 2026, on FTC Act preemption of state AI laws [36].
OMB Memoranda M-24-10 and M-24-18, the operational backbone of Biden-era federal AI use and procurement, were directed to be revised within 60 days under EO 14179. Revised guidance issued in 2025 dropped equity-impact assessments and shifted procurement toward the truth-seeking and ideological neutrality criteria of EO 14319. Most federal agencies retained their Chief AI Officer positions, and the AI Action Plan formalized the Chief AI Officer Council as the primary cross-agency coordination body [21][26].
The AI industry's response has been substantive. In February 2025, OSTP issued a Request for Information seeking input on the AI Action Plan; major labs submitted detailed comments before the March 15, 2025 deadline [37][38].
| Company | Selected positions in OSTP RFI submission |
|---|---|
| OpenAI | Asked the federal government to preempt state AI laws via a narrow voluntary framework; called for faster cybersecurity approvals of AI tools for federal use; proposed public-private partnerships for national-security models |
| Anthropic | Called for streamlining transmission line permitting; proposed adding 50 gigawatts of US energy capacity dedicated to AI by 2027; called for strong NIST and CAISI evaluation roles; supported speeding up federal procurement of AI |
| Microsoft | Asked for expansion of countries qualifying for Tier 1 status under the (then still pending) AI diffusion rule |
| Supported infrastructure permitting reform and federal investment in AI research; raised concerns about export control overreach affecting allied markets |
Most of the labs' procurement and compute concerns were reflected in the final AI Action Plan, including the open-source language, the data center permitting EO, and expanded NSF compute access. The labs differed sharply on the December 2025 state-preemption order. Anthropic, in particular, has at various points endorsed state-level frontier AI legislation, which put it at odds with the administration's preemption push and contributed to a public dispute over Pentagon AI contracting in spring 2026 [39][40].
Civil rights organizations have been broadly critical. Following the December 11, 2025 preemption order, the Lawyers' Committee for Civil Rights Under Law and a coalition of partner organizations called it an "improper attempt to expand presidential authority and allow rampant adoption of unregulated AI" while "trying to neuter the ability of states to protect people, disproportionately Black people and other people of color, who continue to be harmed by discriminatory, unsafe, and biased AI products" [41].
Congressional response has been split along partisan lines but with notable cross-party agreement on rejecting outright preemption. The House version of the One Big Beautiful Bill Act (OBBBA), passed May 22, 2025, originally included a 10-year moratorium on state AI laws. On July 1, 2025, the Senate stripped the moratorium 99 to 1 in an amendment offered by Senator Marsha Blackburn (R-Tenn.) after a coalition of state attorneys general, governors, and consumer advocates objected. Trump signed OBBBA without the moratorium on July 4, 2025 [42][43].
Senator Chuck Schumer's earlier SAFE Innovation Framework, launched in June 2023 with the AI Insight Forums series, did not produce comprehensive legislation. Bills on more targeted AI issues (deepfake nonconsensual imagery, AI in elections, AI labeling, child safety) have advanced in committee but as of mid-2026 no comprehensive federal AI law has passed Congress [44].
As of mid-2026, no federal court has struck down EO 14179 or its follow-on AI executive orders directly. The legal contest is expected to occur primarily through the December 2025 preemption mechanism: when the DOJ AI Litigation Task Force, operational since January 10, 2026, or private parties sue states under the new EO's theories, those cases will produce the first major court rulings on the constitutional limits of the administration's AI policy. State attorneys general in California, New York, Colorado, and Texas have signaled they will defend their AI statutes [29][41].
Despite its relatively brief effective period, Executive Order 14110 had lasting effects on the AI industry.
The order's emphasis on red-teaming, safety testing, and reporting helped normalize these practices across the industry. By the time the order was revoked, most major AI labs had established formal safety testing programs, published safety frameworks, and begun internal governance structures. While these were no longer federally mandated after January 2025, market pressure, customer expectations, and international requirements (particularly from the EU AI Act) kept many companies on a similar trajectory. Voluntary CAISI testing agreements with major labs in 2026, and the voluntary 30-day pre-release review introduced by the June 2026 executive order, indicate that pre-deployment evaluation has continued in modified form even under the deregulatory framework [8][34][45].
The executive order served as a reference point for other countries developing their own AI governance approaches. Its definitions of "dual-use foundation model" and its compute-based threshold for regulatory attention influenced policy discussions worldwide. Even after its revocation, international policy documents continued to reference its frameworks [7].
The cycle of enactment and revocation created significant uncertainty for the AI industry. Companies that had invested in compliance infrastructure for the Biden order found those investments stranded when it was revoked. At the same time, the Trump administration's deregulatory stance was itself unstable, as companies recognized that a future administration could reimpose requirements. This uncertainty complicated long-term planning for AI safety investments [10].
The revocation of the federal order accelerated state-level AI regulation. In the absence of comprehensive federal legislation, states including California, Texas, Colorado, and New York passed their own AI laws in 2025. California's Transparency in Frontier Artificial Intelligence Act (TFAIA, signed by Governor Newsom in September 2025) and Texas's Responsible Artificial Intelligence Governance Act (TRAIGA) both took effect on January 1, 2026. This proliferation of state laws created the very regulatory fragmentation that both industry and the Trump administration sought to avoid, ultimately prompting the December 2025 executive order on federal AI policy preemption [10].
The combination of EO 14156 (energy emergency), EO 14318 (data center permitting), and the AI Action Plan's infrastructure pillar produced one of the largest planned data center buildouts in US history. By late 2025, the Department of Energy had selected four federal sites for data center development, the Air Force had solicited proposals for five additional bases, and Stargate had announced expansion to multiple new sites. Capital expenditure on US AI infrastructure by hyperscalers and Stargate partners reached unprecedented levels in 2025 and 2026, though some projects, including parts of Stargate, faced reported delays related to commercial disputes among the partners [17][28].
As of mid-2026, the United States lacks comprehensive federal AI legislation. The policy landscape is characterized by the Trump administration's pro-innovation, deregulatory stance at the federal level, coupled with an expanding patchwork of state AI laws that the administration is actively trying to preempt through the December 2025 executive order and the DOJ AI Litigation Task Force. The NIST AI Safety Institute has been renamed the Center for AI Standards and Innovation (CAISI) and reoriented toward national security and competitiveness rather than broad AI safety; following the June 2, 2026 executive order it continues to evaluate frontier models but no longer publishes its findings, with review routed to a classified national security process. The NIST AI Risk Management Framework continues to serve as a voluntary industry reference, in a revised form that drops references to misinformation, DEI, and climate change [9][21][45].
Congress has introduced various AI-related bills but has not passed comprehensive legislation. The AI industry operates under a mix of voluntary commitments (such as the Seoul Frontier AI Safety Commitments), state laws, the voluntary 30-day frontier review, and international obligations (particularly for companies operating in the EU). The gap between the regulatory approaches of the US and EU has widened, with the EU AI Act's requirements taking effect in phases through 2026 while the US federal government has moved in the opposite direction [10][45].
The legacy of Executive Order 14110 persists in the safety practices, governance structures, and institutional frameworks it helped establish, even though the order itself is no longer in effect. Whether the Trump framework will produce durable institutional change, or itself be rolled back by a future administration the way EO 14110 was, is the key open question of US AI governance.