Roxanne Heston
Many factors influence where U.S. tech multinational corporations decide to conduct their global artificial intelligence research and development (R&D). Company AI labs are spread all over the world, especially in North America, Europe and Asia. But in contrast to AI labs, most company AI staff remain concentrated in the United States. Roxanne Heston and Remco Zwetsloot explain where these companies conduct AI R&D, why they select particular locations, and how they establish their presence there. The report is accompanied by a new open-source dataset of more than 60 AI R&D labs run by these companies worldwide.Download Full Report
This paper presents new data on the global distribution of U.S. tech companies’ artificial intelligence labs and staff. It focuses on six companies—Amazon, Apple, Facebook, Google, IBM, and Microsoft—all of which have a history of conducting cutting-edge AI research and development. Our findings shed light on where these companies conduct AI R&D and why they select particular locations. This paper also addresses debates about the benefits and risks to national security of having U.S. companies conduct R&D abroad. On its own, new data will not solve these debates, but data is a prerequisite for nuanced policy discussions and cost-benefit analyses. To facilitate future research on the topic, we are releasing our new global AI labs dataset on CSET’s GitHub page.
Key findings include:
Company AI labs are spread all over the world, especially in North America, Europe, and Asia.
Media coverage frequently mentions “AI labs,” but not all companies conduct R&D along a “lab” structure. Amazon, for example, integrates its R&D with the work of its product teams and has no standalone labs. Apple, alternatively, does not release information about its R&D activities.
For the four companies where we could find information on labs—Facebook, Google, IBM, and Microsoft—we found 62 labs conducting AI R&D. The majority of these labs (68 percent) were located outside of the United States.
European countries, mainly the United Kingdom and France, host 19 percent of AI labs in our dataset. Other common destinations were Israel and China (10 percent each) and India (8 percent). The only region in the world without any AI labs was Latin America.
In contrast to AI labs, most company AI staff remain concentrated in the United States.
Whereas most AI labs are located abroad, 68 percent of AI staff at these companies are located within the United States. The San Francisco Bay Area and the Seattle region—where the headquarters of five out of the six companies we focus on are located—host nearly half of these companies’ global AI staff.
The companies combined have 500-plus AI staff in 30 locations across 15 countries. The main company hubs are almost exclusively in large cities, both at home and abroad.
Companies mainly focus on AI research in North America and Europe, and on AI development in other regions.
About 85 percent of labs in North America and Europe do some form of basic AI research, and less than 15 percent focus on development. In contrast, a large majority of labs in other countries, such as India and Israel, focus on applied research or development.
Some companies’ labs appear to focus on research, whereas others focus on development. Most labs conduct multiple types of R&D.
Companies decide to conduct AI R&D abroad for multiple reasons, but access to global talent appears to be the dominant factor.
According to company statements and industry analysts, the main reason that companies set up AI R&D labs abroad is access to talent. Other reasons include cost savings, as well as market access and product adaptation.
Media reports and economic studies suggest U.S. immigration restrictions are associated with U.S. companies’ global expansion. Our findings similarly suggest that immigration restrictions push companies to set up AI labs abroad. At the same time, overseas expansion is unlikely to stop entirely even if U.S. immigration policy became less restrictive.
No comments:
Post a Comment