The White House’s recently released AI Action Plan aims to secure U.S. dominance in artificial intelligence by abolishing red tape and fostering innovation, AI adoption, infrastructure building, and domestic chip manufacturing. To equip and develop the future AI workforce, it pledges public investments in research labs, manufacturing, and technical training programs, as well as tax incentives for businesses to upskill workers and funding cuts to states whose laws unduly hamper AI.
These policies share the commendable goal of creating jobs and uplifting U.S. workers’ living standards, while ensuring that American innovation leads the world. But they’ll be less effective in meeting these objectives without fixing the nation’s broken immigration system, whose deep-seated flaws favor neither America’s economy nor its workers.
Attracting the world’s best and brightest helped the United States to dominate both the space race and the dot-com revolution, while creating unparalleled opportunities and quality jobs for the native-born. The current immigration system, in contrast, threatens the nation’s advantages in the AI race with China by combining the worst of all worlds. It makes it harder for entrepreneurs, researchers, and high-skilled AI talent to stay and work in the United States, while favoring the very staffing firms whose business models are built on undercutting U.S. wages.
Immigrants either founded or co-founded six of the top 10 top U.S. AI companies, and more than half of all the U.S. startups currently valued at more than $1 billion. Indeed, immigrants to the United States start businesses at higher rates than non-immigrants, and have produced 30% of U.S. patents, despite representing just 16% of U.S. inventors. Importantly, 42% of top U.S. AI company founders first came to the country as international students, with foreign students at U.S. universities representing 70% of graduate students in AI-related fields, including up to six in 10 AI-related master’s graduates and more than half of all AI-related PhDs since 2003.
Alas, the United States currently loses 60% of foreign PhD graduates to other countries, despite PhD graduates from top research universities expressing a strong preference to remain in the United States. Roughly half of the departing foreign PhD holders cite immigration issues as motivating their decision, and 60% of foreign PhD holders say they have difficulty staying in the United States, compared to 12% of foreign PhD graduates in other countries. This lost human capital means fewer AI startups and fewer pathbreaking research breakthroughs, meaning less U.S. innovation, less economic growth and fewer jobs and opportunities for Americans. Proposals to abolish the Optional Practical Training (OPT) program that allows international students to work in the United States for a year after they graduate will worsen these problems.
America’s loss is China’s gain. China owns nearly 75% of AI patents worldwide, and the select few universities that are driving Chinese AI chip and quantum research are largely staffed by researchers returning from overseas. The percentage of Chinese students returning to China after completing U.S. doctorates doubled between 2019 and 2022. China has also made strides in attracting top foreign AI talent. Turkish MIT graduate Erdal Arikan has made breakthroughs in AI and 5G telecommunications for Huawei, after leaving the United States for China when he couldn’t secure a U.S. visa. This lost talent not only cedes opportunities to China, but also threatens national security.
These effects imperil billions in US CHIPS Act investment to create a domestic semiconductor-manufacturing industry. Upskilling American workers through training and fixing immigration are complementary objectives. Working alongside foreign talent from nations with established semiconductor-manufacturing sectors allows for crucial tacit knowledge transfers that bolster (and cannot be substituted with) explicit knowledge gained through education. In 2023, Taiwanese semiconductor giant TSMC faced delays in initiating U.S.-based chip manufacturing due to difficulties in securing visas for experienced Taiwanese workers who complement their American counterparts. Since China produces nearly half of the world’s top AI researchers, more opportunities to work alongside this talent would also benefit American researchers, while China loses out.
Foreign talent loss also impacts opportunities for Americans in external and non-AI-related sectors. U.S. cities with more denials of skilled-work visas also see declines in computer-science-related employment for U.S. citizens, and reduced U.S. college graduate wage growth in the following years. The Wharton Business School reports that “policies aimed at reducing immigration have the unintended consequence of encouraging firms to offshore jobs abroad.”
Leaving these findings aside, critics of the H1-B skilled-immigrant visa program rightly note that an influx of foreign tech workers performing the same jobs as native-born Americans lowers wages for those roles, even if it allows U.S. businesses to provide services at lower cost. This has made skilled immigration a polarizing issue among the president’s supporters, as exemplified by high-profile clashes between his pro-immigration tech allies like one-time supporter Elon Musk and anti-immigration “national conservatives.”
But these concerns are exacerbated by the very system that makes it harder to obtain top foreign talent. H1-B visa caps mean that demand far outstrips supply. The government-created lottery to ration the available spots is routinely gamed by tech giants and foreign staffing firms to prioritize large numbers of applications from lower-wage workers. These take priority from high-salary workers and those recruited for specialized backgrounds in critical fields like AI.
The Trump administration will reportedly soon publish a new rule to replace the lottery with a “weighted selection process.” This would give priority to higher-wage workers, those with higher educational attainment, and those meeting other criteria. Higher weighting could also be given to those who work in critical fields, including AI. This would serve to make the United States more competitive and encourage economic dynamism. It could also raise more revenue for funding domestic training programs and facilitating economywide AI adoption, including through income taxes. Immigrants with college or graduate degrees, especially those who come to America when they’re under 45, tend to be net fiscal contributors of more than $1 million over their lifetimes. These contributions can also help to pay down America’s ever-growing fiscal debt. The benefits would be maximized by abolishing the arbitrary visa cap, or at least revising it annually based on national economic growth.
There are other reforms worth considering. The country-of-birth caps on employment-based green cards means that applicants aren’t prioritized based on the value of their skills or their fiscal contribution to the U.S. economy. This has particular adverse effects on top talent from China, India, and a few other nations. In 2020, an Indian advanced STEM degree holder faced an estimated green-card wait time of 195 years. This lowers the average foreign worker’s annual salary by $11,828 by deterring high-skilled workers from staying, thereby reducing their fiscal contributions. Creating or expanding expedited green-card pathways for higher paid U.S. workers and those in AI-related fields would also help address these issues and better align the immigration system with the AI Action Plan’s goals.
These reforms could win bipartisan support, while increasing, rather than depressing, wages and opportunities for America’s native tech and AI workforce. By shoring up competitiveness and dynamism, they would give the United States an edge over China in a field that’s already radically transforming our economy.
