⚡ Quick Summary
Published by The White House, this action plan sets out the Trump Administration's programme for United States leadership in artificial intelligence. It responds to Executive Order 14179, "Removing Barriers to American Leadership in Artificial Intelligence," and is signed by Michael J. Kratsios, David O. Sacks and Marco A. Rubio.
The plan is built on three pillars: accelerating AI innovation, building American AI infrastructure, and leading in international AI diplomacy and security. Each pillar is divided into themed subsections — from removing "red tape" and promoting open-weight models to permitting reform, grid expansion, export controls and biosecurity — that end in "Recommended Policy Actions" assigned to named bodies such as OSTP, OMB, NIST and its Center for AI Standards and Innovation, DOD, DOC, DHS, DOJ and DOL. Recurrent commitments include revising the NIST AI Risk Management Framework to eliminate references to misinformation, diversity, equity and inclusion and climate change, expanding the National AI Research Resource pilot, exporting the full American AI stack to allies, strengthening compute and semiconductor export controls, and evaluating frontier models for national security risk.
The stated objective is to articulate policy recommendations the Administration can deliver to achieve "global AI dominance."
🧩 What’s Covered
The plan moves through three pillars, each closing with agency-assigned policy actions.
- Introduction: frames a "race" for global AI dominance, cites Executive Order 14179 directing creation of the plan, and names three cross-cutting principles — American workers at the centre of policy, AI free from ideological bias, and prevention of misuse or theft of advanced technology.
- Pillar I — Accelerate AI Innovation: covers removal of regulation, frontier AI and free speech, open-source and open-weight models, enabling AI adoption, empowering workers, next-generation manufacturing, AI-enabled science and scientific datasets, advancing the science of AI, interpretability, control and robustness, an AI evaluations ecosystem, government and Department of Defense adoption, protection of AI innovations, and synthetic media in the legal system.
- Pillar II — Build American AI Infrastructure: covers streamlined permitting (NEPA categorical exclusions, FAST-41, Clean Water Act Section 404), developing the grid, restoring semiconductor manufacturing, high-security data centres for military and intelligence use, the infrastructure workforce, critical infrastructure cybersecurity including an AI Information Sharing and Analysis Center, secure-by-design AI, and federal AI incident response.
- Pillar III — Lead in International AI Diplomacy and Security: covers exporting American AI to allies, countering Chinese influence in international governance bodies, compute and semiconductor export-control enforcement, globally aligned protection measures, evaluation of national security risks in frontier models, and biosecurity screening and nucleic acid synthesis controls.
- Recommended Policy Actions: every subsection ends with bulleted actions that name a lead agency, for example OSTP, OMB, DOC through NIST and CAISI, DOD, DHS, DOJ, DOL, NSF, DOE and DOS.
- Sources cited: footnotes reference Executive Orders, Federal Register notices, U.S. Code provisions and NIST publications, including the AI RMF 1.0 and the "Guardians of Forensic Evidence" deepfake work.
💡 Why it matters?
For organisations that build, procure or assess AI systems, the plan signals the direction of US federal policy: deregulation of domestic AI development, agency-level reliance on evaluations and the NIST AI Risk Management Framework, and strict controls on exporting compute and semiconductor manufacturing technology. Compliance and assurance teams can read it to see which federal bodies are named as leads — CAISI at DOC, NIST, DOD and DHS among them — and which instruments the document connects to its goals, such as the NIST AI RMF, Intelligence Community Directive 505 and the NIST AI Consortium. It makes the national-security framing explicit by tying AI governance to export controls and frontier-model evaluations.
❓ What’s Missing
The document states policy goals and recommendations rather than binding obligations: it sets no legal duties, budget figures or completion dates beyond "near-term execution," and gives no metrics for judging success. It leaves open how the proposed NIST AI Risk Management Framework revision would be carried out and approved, and how evaluation results would feed into regulatory decisions. Privacy and civil liberties appear mainly as protections for federal data, while consumer protection, competition and liability receive little detail. The division of authority between federal action and state AI regulation is addressed only through funding considerations and the FCC review.
👥 Best For
Policy and government-affairs teams tracking US federal AI positions; compliance leads who need to see how Washington expects the NIST AI RMF, evaluations and procurement to evolve; national security and export-control analysts following compute and semiconductor measures; and AI vendors assessing what federal adoption, procurement and workforce initiatives may require of them.
📄 Source Details
The document is America's AI Action Plan, carrying the cover line "Winning the Race" and published by The White House in July 2025. It is signed by Michael J. Kratsios, David O. Sacks and Marco A. Rubio. The PDF runs to 28 pages, numbered to 25, ending with a blank page and a further nearly empty page. No URL for the plan itself is printed; footnotes cite Executive Orders, Federal Register notices, U.S. Code sections and NIST publications. The text of all 28 pages was available for this review.
