Your Town Needs AI Experts, Not Just More GPUs

On or before July 22, the Trump administration will receive an “AI Action Plan” that could fundamentally reshape America’s technological future. To realize the goals of the January 2025 executive order on American AI leadership, this plan must prioritize a national strategy for artificial intelligence (AI) literacy—one that systematically breaks down geographic barriers to AI knowledge and creates pathways for all Americans to participate in the AI economy. Recent research led by former Department of Labor Chief Economist Jennifer Hunt reveals a troubling reality: Communities more than 125 miles from AI hotspots see 17 percent lower growth in AI-related jobs and innovation, creating widening opportunity gaps between coastal tech hubs and the rest of America.

For the U.S. to maintain technological leadership, it must do more than establish AI literacy in a few centers—it demands a comprehensive approach to diffusing AI knowledge across the country, much as the Rural Electrification Administration once transformed America by spreading both electrical infrastructure and practical knowledge to communities far from urban centers. The stakes of this challenge extend beyond economic metrics to the very foundations of American competitiveness in an AI-driven future.

The Current Lack of AI Know-How

The current distribution of AI expertise and understanding in the United States reveals concerning disparities. A growing body of research—including Hunt and colleagues’ work—demonstrates that there is a stark geographic concentration of AI knowledge in preexisting innovation hubs. The “superstar” cities—such as San Francisco, Seattle, and New York—of the last wave of technological innovation are doubling down on their talent and infrastructure to thrive in the AI era. While their success is laudable, the fact that the rest of America is missing out raises concerns. “AI innovation hotspots,” to borrow Hunt and colleagues’ term, have disproportionately high rates of AI hiring, growth in AI research and development, and overall economic growth. Those economic and knowledge gains, though, are not spreading. A lack of digital skills and knowledge persists in communities with lower educational attainment, less economic security, and more rural geographic settings.

These geographic disparities in AI expertise manifest in several ways. Large enterprises, particularly those near AI innovation centers, are rapidly incorporating AI capabilities. Hunt and colleagues’ work, for example, demonstrates that finance and insurance industries, traditionally concentrated in major metropolitan areas, are seeing some of the fastest growth in AI job postings. Meanwhile, businesses in sectors such as retail trade and warehousing, often located away from tech hubs, show markedly slower AI adoption rates—as do smaller businesses. The differences in AI uptake among different industries mean that while one region’s workforce upskills on a daily basis, other regions will see their laborers struggle to keep up with new AI advances. Workers in the latter communities both now and for the foreseeable future will find themselves on the wrong side of the tech boom. As time passes, their counterparts in AI hotspots will accumulate ever more knowledge as to the risks and benefits of AI.

The education sector presents another troubling dimension of this knowledge gap. While universities in AI hotspots are rapidly expanding their AI curricula and research programs to cater to local employers, educational institutions outside of these areas most often lack the faculty expertise or resources to offer meaningful AI education. Hunt and colleagues’ finding that distance from AI research centers significantly impacts the growth of AI-related educational opportunities suggests a self-reinforcing cycle of knowledge concentration.

Perhaps most concerning is the trend of leading AI researchers gravitating toward private industry rather than public-facing roles. Hunt and colleagues’ research notes that companies in AI hotspots can offer significantly higher compensation, which ultimately draws talent away from positions that might help spread AI knowledge more broadly. Industry organizations are also able to provide more access to computing power, and the proportion of industry-affiliated research has only risen with the dominance of compute-demanding large language models. This brain drain from public institutions further concentrates AI expertise in select private companies and locations.

Why Diffusing Emerging Technology Knowledge Is Essential to Technological Progress

Historical precedents demonstrate that broad diffusion of technological knowledge is crucial for maximizing societal benefits. The electrification of America offers an instructive parallel. While initial electrical innovations were concentrated in urban centers, President Franklin Roosevelt’s Rural Electrification Administration recognized that spreading both electrical infrastructure and knowledge about its applications was essential for national development. This dual focus on physical infrastructure and knowledge diffusion transformed American agriculture and rural industry. By supporting local cooperatives and private utilities in expanding service to rural communities, the administration helped double the number of rural farms with electricity within just five years.

The internet’s development provides a more recent example. While early internet innovations emerged from concentrated research centers, deliberate efforts to spread both internet access and digital literacy were crucial for realizing the technology’s full potential. The creation of the National Research and Education Network by President George H.W. Bush and subsequent educational initiatives helped ensure that internet knowledge wasn’t confined to technical specialists.

Hunt and colleagues’ research supports this historical pattern. They find that regions with greater AI knowledge diffusion, as measured by the presence of both research and practical applications, show faster overall AI adoption and innovation. This suggests that concentrating AI expertise in a few centers may actually slow down broader technological progress.

Moreover, broader technological understanding supports more informed public discourse and policymaking. When the public has a basic grasp of a technology’s capabilities and limitations, policy discussions can focus on substantive issues rather than misconceptions. This is particularly crucial for AI, where public misconceptions could lead to either unwarranted fear or dangerous overconfidence.

Diffusion of AI Knowledge and America’s AI Dominance

Hunt and colleagues’ research suggests that America’s continued AI leadership requires more than just concentrated excellence in a few centers—especially as allies and competitor nations are racing ahead on promoting the AI knowledge diffusion. China’s AI strategy has long emphasized a need to “comprehensively improve the level of the whole society on the application of AI.” Just last year, Chinese Premier Li Qiang announced an “AI Plus” initiative as part of an effort to “fully integrate digital technology into the real economy” and said that AI is being incorporated at all levels of Chinese public education. Similarly, Russian President Vladimir Putin has stated that “the most important issue [facing Russia] is whether our society and people are ready for the pervasive implementation of [AI],” calling for “mass digital literacy and re-training programmes.” Countries including France, the U.K., Saudi Arabia, Singapore, Malaysia, and Canada have recognized the importance of or are actively promoting national AI literacy. Maintaining competitiveness in AI will require that the U.S. does the same.

As an aside, which of the aforementioned national efforts merits emulating is unclear. The U.K. spun up its program in early 2025. France’s program started in 2018 and has already managed to significantly increase the number of computer science graduates. The CanCode program in Canada, established in 2017, has likewise documented a substantial uptick in AI knowledge among its hundreds of thousands of young participants. A commitment to AI literacy in the U.S. may involve borrowing parts of these and other programs.

Preventing uneven adoption rates is crucial for maintaining public support for AI development. If some regions are indeed falling significantly behind in AI adoption, then the country is at risk of creating “AI deserts” where lack of exposure breeds resistance to the technology. This could lead to political pushback against AI development, similar to historical reactions against other unevenly distributed technological changes.

More broadly distributed AI knowledge would also help address workforce adaptation challenges. Regions with better access to AI knowledge see higher growth in jobs adapting AI to new uses, rather than just jobs being replaced by AI. This suggests that spreading AI literacy could help more workers transition to AI-enhanced roles rather than being displaced.

Proposed Factors for the AI Action Plan

Drawing on Hunt and colleagues’ findings about geographic barriers to AI knowledge diffusion, the AI Action Plan should consider several time horizons for spreading AI literacy:

Short-term initiatives (1-4 years) should focus on breaking down immediate barriers to AI knowledge diffusion. The research shows that physical distance from AI hotspots significantly impedes knowledge transfer. To counter this, the plan could propose the following:

• Creating a national AI extension service, modeled after agricultural extension programs, to bring AI expertise to underserved areas.
• Establishing regional AI resource centers that can serve as intermediaries between major research hubs and local communities.
• Developing standardized AI literacy curricula that can be deployed rapidly through existing educational institutions.
• Incentivizing AI experts to spend time teaching and consulting in regions currently lacking AI expertise.

Medium-term strategies (5-10 years) should focus on building sustainable AI education capacity across the country. Hunt and colleagues’ finding that distance reduces growth in both research and practical AI applications suggests the need for developing local expertise. Key considerations include:

• Supporting the development of AI faculty at regional universities and community colleges.
• Creating partnerships between AI hotspot institutions and more distant educational institutions.
• Developing specialized AI programs tailored to regional economic needs.
• Establishing AI research facilities in currently underserved regions to create new knowledge hubs.

Long-term initiatives (10+ years) should aim to create an ecosystem for continuous AI learning and adaptation. Hunt and colleagues’ finding that state borders impede knowledge flow suggests the need for stronger national infrastructure for AI knowledge sharing. Components might include:

• Building a national platform for AI continuing education.
• Creating mechanisms for sharing AI expertise between industries and regions.
• Developing standards for AI education and certification.
• Establishing frameworks for public-private partnerships in AI education.

The plan should distinguish between different levels of AI knowledge needed because different industries and occupations require varying levels of AI expertise. General AI literacy—basic understanding of AI capabilities, limitations, and implications—is needed broadly. More specialized knowledge—such as AI system development or domain-specific AI applications—is needed in specific sectors and roles.

***

Geographic disparities in AI knowledge suggest that current approaches to AI development may be suboptimal. While concentrated centers of AI excellence have driven impressive advances, the barriers to knowledge diffusion are limiting broader technological progress and potentially creating societal divisions.

The AI Action Plan offers an opportunity to address these challenges by taking a more comprehensive approach to AI knowledge diffusion. By considering short-, medium-, and long-term initiatives to spread both general and specialized AI knowledge, the plan could help ensure that America’s AI leadership is built on a broader and more sustainable foundation than just a few centers of excellence.

Success in the global AI race will require both maintaining America’s leading edge in advanced AI research and ensuring that AI knowledge and capabilities are distributed widely throughout the country. As Hunt and colleagues’ research demonstrates, distance from AI hotspots currently impedes this distribution. By deliberately addressing these barriers through a comprehensive approach to AI knowledge diffusion, the AI Action Plan could help ensure that America’s AI leadership is both technologically advanced and societally sustainable.

This balanced approach—maintaining concentrated excellence while actively spreading knowledge—may be essential for achieving the executive order’s goal of enhancing America’s AI dominance. It would create a more resilient AI ecosystem less dependent on a few geographic centers and better able to adapt to unexpected technological developments like DeepSeek’s recent advances.

The success of such an approach will require sustained commitment and coordination across government, industry, and academia. However, the potential benefits—from more widespread AI innovation to better-informed public discourse about AI—suggest that such investment in AI knowledge diffusion could be crucial for America’s technological future.