AI Is Powering the US Economy, But Who’s Powering AI?

Artificial intelligence is becoming more important to the U.S. economy, driving productivity, powering consumer services, and boosting competitiveness. As a result, the nation's economic future hinges on its ability to develop and attract top AI talent, particularly researchers. However, the Paulson Institute’s Global AI Talent Tracker 2.0 shows that the United States is losing ground to China in both areas and becoming more dependent on Chinese AI talent. To address this, policymakers should establish a special H-1B visa category for foreign AI researchers and allocate five years of National Science Foundation (NSF) funding for university AI research programs that engage students.

The U.S. is struggling to maintain its share of the top 2 percent (or “elite”) of AI talent produced globally as China's rapid growth closes the gap. In 2019, 35 percent of the globe’s most elite (top 2 percent) AI researchers originated from the United States, while only 10 percent originated from China. Yet, over the next three years, the rate of top AI researchers produced in the U.S. slowed, with its global share dropping 7 percent. Meanwhile, China’s production of top AI researchers accelerated, raising the share of top AI researchers originating from China by 16 percent. (See figure 1.)

Figure 1: The share of elite AI researchers originating from the United States and China in 2019 and 2022

When expanding to the top 20 percent (or “top-tier”) of AI researchers, the United States is no longer the top producer of AI researchers. In 2022, only 18 percent of the globe’s top-tier AI researchers originated from the United States, ranking it second to China, which produced 47 percent. More concerningly, the United States is not producing AI talent at a rate that is fast enough to catch up. Indeed, from 2019 to 2022, the global share of top-tier AI talent originating from the United States declined by 2 percent, while China’s share rose by 18 percent. (See figure 2.)

Figure 2: The share of top-tier AI researchers originating from the United States and China in 2019 and 2022

As a result, the United States is becoming more dependent on China for top-tier AI researchers. Indeed, in 2019, while 31 percent of AI researchers working in U.S. institutions were from the United States, 27 percent were from China. Despite a 6 percent growth in AI researchers from the United States working in U.S. institutions from 2019 to 2022, the United States nevertheless saw an increased reliance on Chinese AI researchers. By 2022, the United States relied more on Chinese AI researchers, with 38 percent working for U.S. institutions, compared to 37 percent of U.S. AI researchers. (See figure 3.)

Figure 3: The United States and China’s top-tier AI researchers working in US institutions in 2019 and 2022

Although the United States is increasingly relying on China for top AI researchers, it is becoming a less attractive work destination for AI researchers. In 2019, 65 percent of the most elite (top 2 percent) AI researchers worked in the United States. Yet, by 2022, the share of elite AI researchers working in the United States had fallen to 57 percent. In comparison, this rate rose by 2 percent in the UK and 4 percent in China. Canada and France also saw steep declines. (See figure 4.)

Figure 4: Leading countries where the most elite AI researchers work in 2019 to 2022 (numbers for China’s share in 2019 is below 8 percent but no specific number was given)

When expanding to the top 20 percent (or “top-tier”) of AI researchers, the United States has become an even less attractive work destination. In 2019, 59 percent of the world’s top-tier AI researchers—accounting for both foreign and domestic researchers—chose to work in the United States. Yet, by 2022, it only attracted 42 percent. In comparison, South Korea, Europe, the United Kingdom, and China are becoming increasingly attractive work destinations for these AI researchers. In particular, China saw a 17 percent increase in the share of top-tier AI researchers attracted globally. (See figure 5.)

Figure 5: Leading countries where top-tier AI researchers work in 2019 and 2022

The growth in the number of top AI researchers originating from the United States is slowing compared to China, resulting in the United States’ increased dependence on Chinese AI researchers. Yet, despite this greater dependence, the United States is becoming a less attractive work destination for top AI researchers.

This trend is concerning because advanced AI nations are now in a largely zero-sum contest for AI talent. For every Chinese-born AI researcher the United States attracts, China has one less. Yet, many in the United States view these researchers with suspicion, fearing they could undermine national security and steal intellectual property. Indeed, as the New York Post wrote, “Researchers deputized on behalf of Beijing ended up stealing intellectual property and other technological innovations for batteries, electric vehicles, and semiconductors and are even ‘trying to steal the future of quantum and fusion.’” Although these are legitimate concerns, the solution isn’t to alienate Chinese or foreign researchers from the U.S. AI community but to implement better security measures. This would ensure that U.S. IP and national security are protected without compromising the much-needed AI talent from China and abroad.

Moreover, some are concerned that Chinese and foreign AI researchers working in the United States are taking jobs away from Americans. But these concerns are unfounded. While AI-related jobs are surging, the United States lacks sufficient domestic researchers to fill these roles. Indeed, according to a forecast by Bain & Company, one in two AI-related jobs could remain unfilled by 2027, as job demand is expected to increase to 1.3 million over the next two years while supply is only expected to increase to 645,000. As such, Chinese and foreign AI researchers are not taking jobs away from Americans; instead, they are helping the United States fill a widening AI talent gap.

Policymakers should make it easier for foreign AI researchers with master’s degrees and beyond to work in the United States. This could entail establishing a special H-1B visa category for foreign AI researchers or more proactively establishing a green card system for these researchers. Both options would reduce the challenges foreign AI researchers face in obtaining U.S. work authorizations, attracting more of these researchers to the United States.

Second, Congress must do more to support the education of AI students with advanced degrees in the United States. One challenge is that some leading AI academics have transitioned to the private sector in search of better opportunities. As such, Congress should fund an NSF initiative to provide five-year grants to leading university researchers in AI. This would encourage more students to focus on AI development, boosting the growth of AI researchers in the United States. Failing to take these steps to secure top AI talent will threaten U.S. economic leadership and technological competitiveness.

With the Trump administration adding to the uncertainty for international students here on visas and their efforts to cut federal support for science, these proposals swim against that tide. However, this does not mean that such steps are not important for the United States to keep its lead in AI.