With all the talk about competition between the U.S. and China to dominate AI advancement, it turns out researchers in both countries work together. A lot.
A new report from Stanford University’s Institute for Human-Centered Artificial Intelligence on the global state of AI research and investment states that despite the fact that many AI researchers work together across the globe, the largest number of cross-country AI research and development collaborations between 2010 and 2021 were among people from the U.S. and China working together.
“Even with COVID and issues with respect to collaboration between countries as affected by the previous administration, collaboration between institutions in different countries seems to have continued,” Ray Perrault, co-chair of Stanford’s AI index steering committee and a computer scientist at SRI International, told Protocol last week. “Not much indicates that either of those phenomena had a major dent in collaboration across countries.”
To determine what AI cross-pollination looks like, researchers at Stanford evaluated AI-related research publications such as academic papers and industry-related research.
They noted that AI work between people in the U.S. and China, as represented by published research, far surpassed that of collaborations between the U.S. and other countries, as well as among other countries excluding the U.S. In fact, the Stanford researchers found that U.S.-China AI collaborations produced 2.7 times more publications than produced by AI work between the U.K. and China, the second most prolific partner group in AI research.
Still, after rising steadily fivefold since 2010, the number of U.S.-China AI research partnerships dipped between 2020 and 2021 from more than 10,000 publications to 9,660.
Other top AI partnerships ranked in order according to the number of research publications they produced emerged from teams in the U.S. and U.K.; China and Australia; and the U.S. and Germany, Canada, Australia and France, respectively.