Executive Summary

Talent is core to U.S. competitiveness in artificial intelligence, and international graduate students are a large source of AI talent for the United States. More than half of the AI workforce in the United States was born abroad, as were around two-thirds of current graduate students in AI-related fields. Tens of thousands of international students get AI-related degrees at U.S. universities every year. Retaining them, and ensuring a steady future talent inflow, is among the most important things the United States can do to address persistent domestic AI workforce shortages and to remain the global leader in AI.

This paper holds both good news and bad news for the United States. The good news is that student retention has historically been a core U.S. strength, with well over 80 percent of international U.S.-trained AI PhDs staying in the country, including those from AI competitors such as China. By contrast, other studies have found that the vast majority of China-trained AI talent currently lives outside China. Moreover, contrary to popular perception and anecdotal reports, there is no evidence of recent declines in U.S. retention rates.

The bad news is that two trends are placing this U.S. strength in student retention at risk. The immigration obstacles international graduates face have grown steadily in the past two decades and have worsened in recent years. At the same time, other countries are investing heavily in AI talent attraction and retention, pumping money into their domestic AI ecosystems and opening up their immigration systems to foreign AI talent. In the past, the United States could rely on its status as the world’s sole science and technology superpower to compensate for the flaws of its immigration system, but in today’s more competitive world, complacency is likely to come at a higher cost. Without serious immigration policy changes, the United States stands to lose a vital asset in the international competition for AI leadership.

Results presented below are based on CSET-collected comprehensive career data on 2,000 recent AI PhD graduates from U.S. universities, as well as original analysis of 43,000 immigration records of AI professionals and multiple AI-related survey instruments. Key findings include the following:

• International students are a key source for graduate-level U.S. talent in AI.
  • Two-thirds of graduate students in AI-related programs are international students, and the number of domestic graduate students in these programs has not increased since 1990. Currently, U.S. universities graduate around 50,000 international graduate students (44,000 master’s, 3,000 PhDs) in AI-related fields per year.
  • About 70 percent of immigrants sponsored by AI companies for permanent residency studied at U.S. universities, as did more than half of all international AI workers entering the U.S. labor market each year.
  • International graduates fill critical AI talent gaps in the U.S. labor market. Objective labor market indicators and expert assessments suggest demand for AI talent will far outstrip supply for the foreseeable future.
• Stay rates among international graduates in AI are persistently high.
  • Around 90 percent of international AI PhD students take a job in the United States after graduating, and more than 80 percent stay in the country for at least five years. Past studies strongly suggest stay rates are likely to be high beyond the five-year window for which there is hard AI-specific data.
  • Multiple data sources indicate retention rates have not fallen in recent years, contrary to popular perception and anecdotal reports.
  • Stay rates are highest—exceeding 90 percent—among students from Taiwan, India, Iran, and China, and lower—around 75 percent— among students from European countries.
  • Among the few graduates who leave the United States, the large majority go to U.S. allies and partners in Europe and Asia, such as the U.K., Canada, Singapore, and South Korea. Less than 20 percent of those leaving go to China.
• Professional considerations are the main reasons for international talent to stay in the United States, while immigration difficulties and cultural factors are the most important issues pushing away talent.
  • The U.S. private sector is especially attractive to graduates; around 60 percent go on to work for companies after completing their degree, with most of the remainder going into academia.
  • Graduates with ambitions to launch or work at startups are particularly hampered by immigration obstacles. Whereas more than 40 percent of domestic graduates who go into the private sector work at small companies, less than 20 percent of international graduates do so.

On the policy font, research highlights two important trends that, together, could erode the U.S. AI talent advantage:

• Domestically, international graduates who want to stay are faced with significant obstacles in the U.S. immigration system, and these problems are getting worse.
  • Green card wait times have increased significantly in recent years. One study estimates that an Indian AI PhD graduate sponsored for a green card today would face a wait time of around 50 years in the absence of immigration reforms.
  • Optional Practical Training, a program used by tens of thousands of international graduates from AI-related programs every year, is currently facing significant legal and policy challenges. Given the lack of available alternative visas for these graduates, many would likely be forced to leave the United States if OPT were eliminated.
  • There is no suitable U.S. entrepreneur visa for international graduates who want to start AI companies. Sponsoring employees for visas is often too costly for startups, in large part due to inflexible and long application timelines.
• Internationally, the United States faces increasing competition for top AI talent.
  • The United States has lost its historical near-monopoly on AI R&D and commercial activity. In 2013, the United States accounted for more than 70 percent of funding deals for AI startups. By 2018, this number had dropped to 40 percent.
  • Other countries are opening their immigration systems and aggressively recruiting U.S.-trained AI talent. Nearly two dozen countries have recently launched startup visa programs marketed mainly to tech entrepreneurs.
  • Other countries are also investing heavily in their education systems. The number of U.S. universities reporting international students declining admission offers because they preferred to study at home or in third countries increased three-fold between 2016 and 2018.

Based on these findings, the report lays out two priorities and several concrete options for U.S. policymakers:

• First, policymakers need to reform high-skill immigration rules in order to maintain and improve U.S. international AI talent retention. Options for achieving this include:
  • Reforming student visa regulations and procedures, for example by codifying OPT in statute and eliminating processing backlogs.
  • Streamlining post-graduation transitions into the U.S. labor force, as could be done through the creation of a statutory student-to-work pathway and a dedicated visa program for entrepreneurs.
  • Shortening the path to permanent residency and citizenship, for example by removing numerical caps for in-demand graduate talent or creating accelerated citizenship-through-service programs.
• Second, policymakers should address legitimate security concerns around foreign AI talent while avoiding broad and potentially counterproductive restrictions. This can be done by:
  • Improving policy coordination domestically and internationally by creating a new interagency task force and increasing engagement with allies, without whom counter-transfer efforts for diffuse technologies such as AI would almost certainly be ineffective.
  • Raising awareness of transfer practices through open-source collection and dissemination, for example by allocating more resources to open-source intelligence activities or adopting FARA-like legislation for foreign talent recruitment activities.
  • Collecting more and better data about student retention trends, including among master’s students, for whom there is no government survey or other data source that tracks post-graduation career choices.