Skilled and Mobile: Survey Evidence of Immigration Preferences of AI Researchers.

Remco Zwetsloot, Baobao Zhang, Noemi Dreksler, Lauren Kahn, Markus Anderljung, Allan Dafoe, Michael C. Horowitz

Abstract

Countries, companies, and universities are increasingly competing over top-tier artificial intelligence (AI) researchers. Where are these researchers likely to immigrate and what affects their immigration decisions? We conducted a survey (n = 524) of the immigration preferences and motivations of researchers that had papers accepted at one of two prestigious AI conferences: NeurIPS and ICML. We find that the U.S. is the most popular destination for AI researchers, followed by the U.K., Canada, Switzerland, and France. A country's professional opportunities stood out as the most common factor that influences immigration decisions, followed by lifestyle and culture, the political climate, and personal relations. Visa and immigration difficulties were perceived to be a particular impediment to conducting AI research in the U.S., the U.K., and Canada.

Introduction: The Global Competition for AI Talent

The global "war for talent" in artificial intelligence is intensifying as nations recognize AI as a pillar of future economic growth and national security. However, most existing data on where AI researchers move comes from "inferred" sources—like tracking institutional affiliations on published papers. This research addresses the lack of direct evidence regarding the motivations and preferences of these researchers. By surveying the researchers themselves, we aim to understand what makes a country attractive, what drives the decision to move, and whether current immigration policies act as a barrier to innovation.

Methodology

The study utilized a pre-registered survey of 524 researchers who had papers accepted at top-tier conferences: the Conference on Neural Information Processing Systems (NeurIPS) and the International Conference on Machine Learning (ICML).

Key methodological features include:

  • Sampling Frame: Authors from NeurIPS 2018, ICML 2018, and a 2016 expert survey.
  • Citizenship Proxy: The country where a respondent completed their undergraduate degree was used as a proxy for their country of origin.
  • Regional Analysis: Respondents were grouped by current residence (Asia, Europe, North America) to analyze how geography influences priorities.

Key Insights and Drivers

Professional Gravity

Professional opportunities and environment are the primary drivers for relocation. A staggering 91% of researchers cited these factors as top considerations. For elite talent, the choice of destination is largely dictated by the quality of colleagues, labs, and research challenges available.

The U.S. Paradox

While the U.S. remains the most popular destination, it is also perceived as having the most problematic immigration system. This creates a tension between the desire to work in the world's leading AI hub and the practical difficulties of securing the necessary legal status to do so.

Policy as a "Tie-Breaker"

While professional environment is the leading factor, nearly half (47%) of researchers noted that ease of immigration or specific incentives influenced their decisions. This suggests that while policy cannot compensate for a lack of opportunities, it can act as a decisive factor when choosing between comparable offers.

Key Results: Destinations and Deterrents

Top Destinations

When asked which countries they would consider moving to, the United States was the clear leader, followed by several European and North American hubs.

Figure 1: Percentage of respondents considering various countries for potential relocation.

As shown in Figure 1, after the U.S. (58%), the most popular destinations for potential relocation were the U.K. (35%), Canada (28%), Switzerland (25%), and France (16%).

Immigration Hurdles

Perceptions of visa difficulties as a "serious problem" for research varied significantly by country. These hurdles are most pronounced in the very countries that are most sought after by researchers.

Figure 2: Percentage of researchers in each country who view immigration/visa issues as a serious problem for conducting research.

As illustrated in Figure 2, 69% of U.S.-based researchers see visa issues as a serious impediment. This is followed by the U.K. (44%) and Canada (29%). In contrast, only 10-18% of researchers in China, Germany, and France viewed their immigration systems as a serious barrier to their work.

Limitations and Future Work

The survey had a 17% response rate, with a slight academic skew. Additionally, the data represents a snapshot from late 2019; subsequent global events, such as the COVID-19 pandemic and shifting geopolitical tensions, may have influenced these preferences. Future research should consider longitudinal studies to track how these preferences evolve over a researcher's career and investigate specific policy interventions that effectively attract and retain top-tier talent.

Summary updated

Abstract

Countries, companies, and universities are increasingly competing over top-tier artificial intelligence (AI) researchers. Where are these researchers likely to immigrate and what affects their immigration decisions? We conducted a survey $(n = 524)$ of the immigration preferences and motivations of researchers that had papers accepted at one of two prestigious AI conferences: the Conference on Neural Information Processing Systems (NeurIPS) and the International Conference on Machine Learning (ICML). We find that the U.S. is the most popular destination for AI researchers, followed by the U.K., Canada, Switzerland, and France. A country's professional opportunities stood out as the most common factor that influences immigration decisions of AI researchers, followed by lifestyle and culture, the political climate, and personal relations. The destination country's immigration policies were important to just under half of the researchers surveyed, while around a quarter noted current immigration difficulties to be a deciding factor. Visa and immigration difficulties were perceived to be a particular impediment to conducting AI research in the U.S., the U.K., and Canada. Implications of the findings for the future of AI talent policies and governance are discussed.
BibTeX
@misc{zwetsloot2021skilled,
  author = {Remco Zwetsloot and Baobao Zhang and Noemi Dreksler and Lauren Kahn and Markus Anderljung and Allan Dafoe and Michael C. Horowitz},
  title = {{Skilled and Mobile: Survey Evidence of Immigration Preferences of AI Researchers.}},
  year = {2021},
  eprint = {2104.07237},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2104.07237},
}
Figure