America’s AI Action Plan: A Heady Cocktail of Techno-optimism, Culture Wars and Geopolitics

On July 23, 2025, the White House unveiled its Artificial Intelligence Action Plan (the “Plan”), following President Donald Trump's executive order in January titled "Removing Barriers to American Leadership in Artificial Intelligence."  

Where previous policy focused on establishing guardrails through executive orders and agency guidance, the new Plan prioritizes deregulation and rapid deployment to accelerate AI adoption across key economic sectors. 

CORE FRAMEWORK AND PRINCIPLES 

The Action Plan operates on three foundational pillars: innovation, infrastructure, and international diplomacy and security.  

PILLAR I: ACCELERATING AI INNOVATION 

Regulatory Environment and Federal Procurement 

To reduce red tape and promote innovation, the Plan outlines several measures: 

• Funding Decisions: Federal agencies will consider a state's AI regulatory climate when making decisions on discretionary funding. States with regulatory regimes deemed unfavourable to AI development may face funding limitations. 

Critics argue that this proposal amounts to a "soft moratorium" on state-level AI regulation, following a failed legislative attempt earlier in 2025, in which a formal 10-year moratorium on new state and local AI laws was rejected by the Senate. By conditioning federal support on regulatory permissiveness, the Plan is said to achieve similar outcomes without explicit pre-emption. 

• Regulatory Review: The Plan proposes a review of Federal Trade Commission investigations conducted under the Biden administration. This includes examining final orders, consent decrees, and injunctions to ensure that liability theories do not impede AI innovation. 

• Procurement Guidelines: New federal procurement guidelines will require government contracts exclusively with large language model developers who ensure their systems are objective and free from ideological bias. 

Landry Signe, senior fellow of Global Economy and Development, argues that the Plan inadequately addresses AI governance issues, including accountability, ethics, and transparency, creating risks of unregulated AI harms and privacy erosion.  

He contends the Plan overlooks the essential role of regulators and public institutions in addressing two critical challenges: the inability of legal frameworks to keep pace with rapidly advancing technology, and the delays and erosion of public trust caused by uncoordinated inter-agency actions. 

Content Standards  

The federal government aims to ensure AI systems "objectively reflect truth rather than social engineering agendas." Proposed actions include revisions to the National Institute of Standards and Technology (NIST) AI Risk Management Framework and conducting research on frontier models from China to evaluate security and censorship issues. 

Aaron Klein, former U.S. Treasury official, argues the Plan's emphasis on objectivity overlooks fundamental challenges in AI development. He contends that "AI operates on existing data, collected and processed over decades of discrimination," asserting that the Plan's focus on innovation fails to address inherent biases in training datasets 

Open-source and open-weight models are designated as having geostrategic significance. The government seeks to foster their development by improving access to large-scale computing power for startups and academics, and by enhancing financial markets for compute resources. 

Industry and Government Operations Adoption 

The Plan recognises several barriers to AI adoption in critical sectors, including distrust, limited technology understanding, and complex regulatory frameworks. Proposed solutions include: 

• Establishing regulatory sandboxes. 

• Facilitating open data sharing. 

• Engaging with domain-specific stakeholders to encourage the adoption of national AI system standards. 

Critics have warned that regulatory sandboxes may have unintended consequences, such as regulatory capture and undermining public trust.  

Ivan Lopez, a Stanford HAI Tech Ethics Policy fellow, cautions that, while the administration seeks to accelerate AI adoption in vital sectors like healthcare, medical professionals are calling for greater caution. They stress the need for rigorous evaluation and oversight, warning that treating healthcare merely as a productivity frontier risks serious consequences, including mis-triaged emergencies, biased predictions, and a breakdown of patient trust. 

To transform government operations, the Plan designates the Chief AI Officer Council as the primary coordination body for federal AI use. It also proposes creating personnel exchange programs to share AI expertise across agencies and establishing centralised AI procurement toolkits. 

Collaboration with American AI developers is proposed to protect innovations from security risks, including cyber threats and insider threats. 

Defence Applications and Advanced Technologies 

The Department of Defence (DoD) will implement AI by identifying required skills, developing training programs, and establishing a Virtual Proving Ground for testing AI and autonomous systems.  

Plans include creating processes to identify and improve high-impact DoD workflows for AI automation. Additional measures include finalising agreements with cloud and technology providers for priority computing access during national emergencies and expanding Senior Military Colleges as centres for AI research and talent development. 

The Plan proposes investment in autonomous technologies, including drones, self-driving vehicles, and robotics for industry and national security applications. It also calls for convening stakeholders to identify supply chain challenges for American robotics and drone manufacturing. 

Scientific infrastructure modernisation efforts include investment in automated cloud-enabled laboratories, using long-term agreements to support focused research, and incentivising researchers to release high-quality datasets publicly. Federally funded researchers will be required to disclose non-proprietary, non-sensitive datasets used in their AI model development and experimentation. 

Frontier AI Research and Synthetic Media Countermeasures 

Frontier AI system research will focus on interpretability and safe deployment in critical defence applications. Proposed actions include technology development programs that advance AI interpretability, control systems, and adversarial robustness, with integration into the forthcoming National AI Research and Development Strategic Plan. 

Synthetic media countermeasures will address malicious deepfake challenges beyond the scope of the TAKE IT DOWN Act. This will be done through the development of NIST's Guardians of Forensic Evidence evaluation program into formal guidelines and voluntary forensic benchmarks. 

PILLAR II: BUILDING AMERICAN AI INFRASTRUCTURE 

Development and Permitting Streamlining 

AI infrastructure, including data centres, semiconductor manufacturing facilities, and energy sources, is often delayed by the current U.S. permitting system. The Plan proposes to streamline this process by: 

• Establishing new categorical exclusions under the National Environmental Policy Act for data centre actions with no significant environmental effects. 

• Expanding the use of the FAST-41 process for data centres and energy projects. 

• Making federal lands available for data centre and power generation infrastructure construction, while maintaining security measures. 

Energy Grid and Semiconductor Manufacturing 

Electric grid modernisation is crucial to meet the rising energy demands of data centres and AI-driven industries. Proposed actions include: 

• Stabilising and improving power grid efficiency. 

• Ensuring nationwide resource adequacy standards are met to prevent shortages. 

• Aligning financial incentives with grid stability and actual system requirements. 

The restoration of semiconductor manufacturing aims to boost domestic job creation and strengthen technology leadership. This includes ensuring a return on taxpayer investment, removing non-essential policy requirements from CHIPS-funded projects, and reviewing semiconductor grants and research programs to ensure they promote advanced AI tool usage. 

Security Infrastructure and Cybersecurity 

The Plan emphasises the need for high-security infrastructure for military applications, protected against advanced nation-state cyber threats. It recommends creating technical standards for high-security AI data centres and advancing the adoption of classified compute environments for scalable, secure AI workloads. 

To address the growing cybersecurity threats posed and defended by AI, the Plan proposes: 

• Establishing an AI Information Sharing and Analysis Centre for threat information sharing across critical infrastructure sectors. 

• Issuing remediation guidance to private sector entities. 

• Facilitating collaborative sharing of known AI vulnerabilities. 

Proposed refinements to address adversarial inputs and system vulnerabilities include updating the DoD's Responsible AI and Generative AI frameworks and publishing Intelligence Community Standards on AI Assurance. 

Incident Response Framework 

To mitigate risks from AI failures, the Plan will embed AI-specific incident response protocols into existing frameworks. This will be accomplished by modifying the Cybersecurity and Infrastructure Security Agency's Cybersecurity Incident & Vulnerability Response Playbooks and encouraging the responsible sharing of AI vulnerability information. 

PILLAR III: INTERNATIONAL AI DIPLOMACY AND SECURITY 

Export Strategy and Alliance Building 

The Plan establishes export programs for comprehensive U.S. AI technology stacks, including hardware, models, software, applications, and standards to allied nations. The administration aims to operationalise programs that gather industry consortium proposals for these full-stack AI export packages. 

Cameron F. Kerry, former general counsel and acting secretary at the U.S. Commerce Department, argues this "all-or-nothing" approach may heighten concerns about technological dependence among potential partners. He contends that promoting complete American technology stacks, rather than enabling flexible collaboration, complicates efforts by other nations to achieve balanced partnerships while avoiding excessive reliance on foreign technology. 

CONCLUSION 

The AI Action Plan represents the Trump administration's vision of American technological leadership through deregulation, rapid deployment, and strategic competition with global rivals, particularly China.  

This approach sharply contrasts with the EU's regulatory framework: while the EU's AI Act of 2023 imposes legally binding obligations with strict penalties, prioritising fundamental rights and transparency through precautionary regulation, the U.S. Plan embraces market-driven innovation and rapid deployment in strategic sectors.  

The Plan seeks to position the United States as the dominant force in AI development by removing regulatory barriers, streamlining infrastructure development, and promoting comprehensive technology exports to allied nations.  

However, policy experts identify fundamental tensions in this approach. The Plan's emphasis on objectivity and innovation may inadequately address inherent algorithmic biases and systemic discrimination, while its deregulatory framework risks undermining public trust and accountability mechanisms. 

The conditioning of federal funding on state regulatory environments raises questions about federal-state relationships and civil rights protections. Moreover, rapid AI adoption in critical sectors such as healthcare and defence, without corresponding rigorous oversight, presents potential risks to public safety and national security.  

Ultimately, the Plan reflects a fundamental policy choice: prioritising competitive advantage and technological supremacy over comprehensive governance frameworks. Whether this approach proves sustainable will depend not only on America's ability to maintain its AI leadership, but on its capacity to address the governance deficits that critics have identified without compromising its innovation objectives.