AI Action Plan and Executive Orders

KPMG Regulatory Insights

Shifting to Greater ‘Self-Regulation’: With a general shift away from a federal rules-based approach, expect a growing federal movement toward industry 'self-regulation’.
Setting Domain-Specific National Standards: For specific domains/industries (e.g., healthcare, energy, agriculture), public-private collaboration in order to develop and adopt domain-specific national standards for AI systems.
Opening the Sandboxes: To drive a “try-first” culture of exploration and adoption, anticipate the quick establishment of agency-sponsored regulatory sandboxes (or AI Centers of Excellence).
Coming Scrutiny and Divergence: Given the rapid development and evolution of AI technologies, expect continued legal actions, jurisdictional regulatory divergence, and an evolution of regulatory frameworks over time.

July 2025

The Administration releases its AI policy recommendations and executive directives, including:

AI Action Plan, which introduces over 90 policy recommendations across three pillars of innovation, infrastructure and diplomacy and security
AI Executive Orders, related to: expediting permits for data center infrastructure; promoting global export of US AI technology; and procuring “unbiased” large language models.

1. AI Action Plan

Executive Order 14179, “Removing Barriers To American Leadership in Artificial Intelligence,” directed the development of an AI Action Plan, issued by 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.

As released, the AI Action Plan puts forth over 90 policy recommendations structured around three pillars, including:

Pillar One: Accelerate AI “Innovation”
Area	Key Recommended Actions, including:	Agencies, including:
Perception of “Onerous Regulation”	Eliminate obstacles to private-sector AI innovation Revise or repeal regulations that may hinder AI development or deployment	OSTP/OMB
State AI Law Rights	Federal government should not interfere with states’ rights to pass AI laws not perceived as “unduly restrictive to innovation” Federal government should not allow AI-related Federal funding to be directed to states that have an “AI regulatory regime” that “may hinder the effectiveness of that funding”	OSTP/OMB, FCC, FTC
Open-Source and Open-Weight AI	Improve financial markets for computing power Increase research community access to private sector resources Publish a new National AI Research and Development Strategic Plan to guide Federal AI investment	DOC/NIST, OSTP, NSF
AI Adoption	Establish regulatory sandboxes for rapid AI deployment and testing Launch domain-specific efforts (e.g., by industry) to accelerate development and adoption of national AI standards and measure AI impacts to productivity	FDA, SEC, DOC/NIST, DOD
Manufacturing and Science	Invest in emerging technologies for autonomous manufacturing systems Develop automated cloud-enabled labs and incentivize the public release of high-quality datasets Create minimum data quality standards for the use of scientific data in AI model training	DOD, DOC, DOE, NSF

Pillar Two: Build Infrastructure
Area	Key Recommended Actions, including:	Agencies, including:
Permitting	Reform NEPA regulations for data center development Use U.S. products in infrastructure development	DOC, DOE
Power Grids	Stabilize and optimize the current power grid system Embrace new energy generation sources (e.g., geothermal, nuclear fission) Change power markets to align financial incentives with grid stability	DOE
U.S. Semiconductor Manufacturing	Remove perceived “burdens” on semiconductor projects	DOC
Data Centers	Build high-security data centers for military and intelligence Establish technical standards for secure AI data centers	DOD, NSC, DOD
Infrastructure Workforce	Identify priority roles for AI-related infrastructure buildout Create industry-driven training programs tied to priority AI infrastructure occupations	DOL, DOC
Cybersecurity	Share AI-security threat information through a new AI-ISAC partnership Publish standards for secure AI systems through DOD and ODNI Update incident response frameworks for AI vulnerabilities and coordinate efforts	DHS, DOD, ODNI

Pillar Three: Lead in Diplomacy and Security
Area	Key Recommended Actions, including:	Agencies, including:
Allies and Partners	Facilitate export of full AI technology stack globally	DOC
Foreign Influence	Advocate for governance approaches/regulations favoring U.S. position	DOC, DOS
Export Controls	Use location verification features to prevent chip diversion Address gaps in semiconductor export controls and develop export controls on semiconductor subsystems	DOC, OSTP, NSC
Protection Measures	Coordinate with allies on complementary AI protection and export controls	DOC, DOS
National Security Risks	Evaluate national security risks for AI frontier systems through public-private partnership with developers	DOC, DOD
Biosecurity	Screen sequences and customers to prevent misuse of biological data Build national security-related AI evaluations	OSTP, DOC

Note: The Assistant to the President for Science and Technology anticipates the recommended policies in the AI Action Plan can be executed in the next six months to one year.

2. Executive Orders

Three new AI-related Executive Orders establish expectations for:

Rapid buildout of AI data centers and related infrastructure, including provisions for federal financial support (e.g., loans, grants, tax incentives) and expedited, streamlined permitting processes and reviews.
A coordinated national effort, including technical, financial, and diplomatic resources, to support the development and global deployment of full-stack AI export packages through a new American AI Exports Program.
Federal procurement of large language models (LLMs) developed in accordance with “unbiased” AI principles, building on EO 13960.