The Myth of China's "AI Talent Pipeline"
Deep Dives
Explore related topics with these Wikipedia articles, rewritten for enjoyable reading:
-
Gaokao
12 min read
The article extensively discusses the Gaokao as a central pressure point in China's education system, but most Western readers lack deep understanding of this examination's structure, history, and societal impact that shapes millions of lives annually
-
International Olympiad in Informatics
13 min read
The article specifically mentions Chinese students outperforming OpenAI at IOI and discusses the Olympiad track as a key pathway in China's talent pipeline. Understanding this competition's format and prestige provides important context
-
996 working hour system
13 min read
The article describes grueling schedules from childhood through academia (6am-10pm days, 3-hour sleep rotations, 11-hour daily PhD requirements). This phenomenon of extreme work culture in China, formalized as '996', provides essential context for understanding the systemic pressures discussed
Zilan Qian is a program associate (research) at the Oxford China Policy Lab and holds a Master’s degree in Social Science of the Internet from the University of Oxford.
Trigger warning: the second half of this article explores suicide.
“The US-China AI race is a race between Chinese — those in the US vs. those in China.”
This joke has real-world references. It is no secret that Chinese engineers and researchers make up a meaningful percentage of the AI workforce in the US. According to the Paulson Institute’s Global AI Talent Tracker 2.0, by 2022, US institutions relied more on Chinese AI researchers (38%) compared to US AI researchers (37%). Yet, this tracker still underestimates the Chinese AI talents in the US, because researchers are only counted as Chinese if their undergraduate degree is from a Chinese institution. That excludes a massive number of China-born AI researchers who did their undergraduate degrees in the US.
Meanwhile, China’s own AI progress, almost 100% powered by China-born Chinese, has grown at an unmatched pace. Besides the industry performance that can compete with the US, in 2024, China’s AI research publication output matched the combined output of the US, UK, and European Union, and now commands more than 40% of global citation attention.
People often cite China’s talent pipeline as one of its most valuable strategic resources — a system to admire or even emulate. Unfortunately, this view is fundamentally wrong. The system is highly inefficient, with a low cost-return rate: the top STEM genius everyone sees at the summit is built upon the bodies of massive numbers of talented students who failed to reach the top.
This piece is not about the life stories of successful Chinese AI or STEM talents. It is not about how the talent system works — but about how it does not. It explores the price paid to create this talent pool and the untold mental health stories behind it, as experienced and witnessed by me.
How to Build an “AI Talent Pipeline”
I grew up in Hangzhou, which is known today as one of China’s booming AI and robotics hubs. I went to some of the city’s top middle and high schools, the kinds of places that sit at the center of the country’s STEM pipeline. A middle school senior several years ahead of me became the co-founder of xAI, and another high school ...
This excerpt is provided for preview purposes. Full article content is available on the original publication.