The White House released its 2025 AI Action Plan—a 28-page blueprint focused on securing US leadership in artificial intelligence. It’s an executive-led strategy with minimal reliance on Congress, emphasizing rapid deployment and national competitiveness.
The plan is organized around three pillars: innovation, infrastructure, and international engagement. Each pillar underscores how AI is not just a technology, but a strategic asset in a new era of global competition.
Accelerate AI Innovation: The plan calls for investments in federal R&D for foundational AI models and expands access to secure testbeds for real-world deployment. It supports open-source and open-weight model development. It also expands AI-related workforce initiatives, including K-12 education, reskilling, and apprenticeships.
Build AI Infrastructure: The plan aims to remove chokepoints in physical infrastructure by fast-tracking permitting for data centers, semiconductor facilities, and energy projects, primarily through expanded exclusions in the National Environmental Policy Act and streamlined environmental review. It also calls for a national strategy to modernize the power grid and embraces next-generation energy sources like nuclear fission, fusion, and enhanced geothermal to meet the soaring demand for AI technologies.
Lead in International AI Engagement: Promotes US-developed AI models, hardware, and standards abroad and tasks federal agencies to engage in diplomatic and standard-setting bodies to “vigorously advocate for international AI governance approaches that promote innovation, reflect American values, and counter authoritarian influence.” It also calls for national security evaluations of emerging AI capabilities, led by the Center for AI Standards and Innovation.
Several key policy actions stand out and merit particular attention:
Open as a National Priority: One of the most striking departures from the Biden administration is the embrace of open-source and open-weight models as a strategic imperative, an issue I wrote about last year. The plan argues that models “founded on American values” should lead the way. This is especially important in the Global South, where open systems will likely be the primary means of AI access. Restricting these models unnecessarily risks ceding influence to Chinese alternatives that may carry embedded authoritarian values and security risks. Meta’s Mark Zuckerberg has argued this point, and OpenAI’s Sam Altman wrote, “The challenge of who will lead on AI is not just about exporting technology, it’s about exporting the values that the technology upholds.”
Regulatory Sandboxes: This is one of the plan’s most underappreciated strengths. Regulatory sandboxes provide a practical approach for researchers, startups, and companies to rapidly deploy and test AI tools in controlled environments, while committing to transparent data sharing and results. These environments not only help regulators build their technical understanding of new technologies but also surface risks before they scale.
One early example is the Coalition for Health AI, which brings together a diverse group of stakeholders to develop best practices and frameworks for the responsible use and implementation of AI in healthcare. The blending of government and industry expertise facilitates the exploration of key questions surrounding the quality of AI tools and develops a shared understanding of these technologies and their uses in different areas of healthcare.
But more experimentation is needed—both in coalition design and in sector focus. Establishing sandboxes in areas like education and child development could offer vital insights into how AI supports tutoring and workforce readiness, while also assessing potential negative effects (particularly AI companions) on social connection, child well-being, and developmental outcomes.
Human Capital: As with energy and compute, AI runs on people: Human capital is the engine of AI competitiveness. The plan encourages federal agencies to prioritize AI skill development across education and workforce funding streams, with a focus on apprenticeships and industry-recognized credentialing programs.
It also highlights the Department of Labor’s AI Workforce Research Hub as a key initiative to evaluate AI’s evolving impact on the labor market. Debate continues over whether AI will supercharge productivity or lead to widespread job displacement. As the impacts of emerging technologies on employment are difficult to forecast, the plan emphasizes the need for improved, near-real-time labor market data to track adoption trends and inform timely policy responses.
The bottom line: This is not a risk-first governance plan. As James Pethokoukis reflected, it is “proactionary rather than precautionary.” It’s a pro-growth, national security-driven roadmap that seeks to accelerate US deployment and strengthen domestic capacity. The challenge ahead will be translating executive ambition into institutional action and doing so in ways that sustain public trust while keeping pace with rapidly evolving AI capabilities.
Below is an audio summary of this blog post:
The post The AI Action Plan: Securing America’s Future in the Age of Intelligence appeared first on American Enterprise Institute – AEI.
