CSET: global AI research more than doubled between 2017 and 2022; Chinese Academy of Sciences was the top producer of highly cited AI papers, followed by Google
Context & Ripple Effects
This extends a long-running split in the coverage: Chinese scholars had already been reported to out-publish U.S. peers, while questions persisted over research quality. The new CSET ranking places the Chinese Academy of Sciences at the top on a highly cited-paper measure, sharpening that quality dimension.
It also complicates a more recent finding that U.S.-based work and Google dominated the most-cited annual papers from 2020 through 2022: Google’s earlier citation lead now sits alongside the Chinese Academy of Sciences’ leading position across CSET’s broader 2017–2022 window. Meanwhile, China’s large researcher pipeline has been linked to heavy investment in AI education in a recent talent study.
First-order effects
- The Chinese Academy of Sciences gains a prominent research-performance benchmark, while Google is identified as the next-leading producer of highly cited AI papers over the same period.
- The reported doubling of global AI research output raises the competitive bar for institutions seeking visibility and influence through published work.
Second-order effects
- The differing results across citation analyses will intensify scrutiny of methodology—such as time window and how influential papers are counted—rather than allowing a single national or organizational ranking to settle the research race.
- Google’s placement alongside a state-backed research institution underscores that frontier research competition is being measured across both corporate labs and national research systems.
Third-order effects
- If research output continues to expand, citation-based rankings are likely to become a more contested proxy for AI capability, with greater weight on which institutions can turn large talent pools into influential work.
- The pattern points toward a more multipolar AI research base, where state-supported institutions and global technology companies compete simultaneously for talent, scientific standing, and downstream technological influence.
The trend: AI research leadership is becoming a contest between scaled national research ecosystems and corporate labs, with output, talent, and citation impact producing different but complementary measures of strength.