Economists debate whether AI will repeat the 'China shock'—a shift in manufacturing jobs to cheaper labor—but this time targeting white-collar work instead of factory floors.

Apollo chief economist Torsten Slok argues the AI shock mirrors the 2001 China WTO entry shock: China's production explosion accounted for 59.3% of all U.S. manufacturing job losses between 2001 and 2019 (about 4 million jobs), yet overall U.S. unemployment remained low and real manufacturing value added rose 50% from 2001 to 2024. Slok sees AI following the same pattern, shifting job displacement from factory floors to cognitive and white-collar work.

Slok invokes Jevons paradox to predict AI will create net job growth: when the Watt steam engine improved coal-fired engine efficiency in the 1800s, coal consumption actually increased as energy became cheaper. Similarly, as AI automates white-collar tasks, the market for those positions expands—and in radiology, the number of active radiologists in the U.S. over the past decade has grown by about 10% despite AI automating parts of the imaging process.

Economist David Autor, who coined the term 'China shock,' disagrees with the parallel. He argues AI will displace jobs differently than the China shock did: AI targets job functions rather than specific industries or regional geographies, making labor changes potentially more widespread. AI productivity gains may appeal to firms but prove more disruptive to labor than the China trade shock, which firms experienced as a pure negative competitive shock.