Surveys last spring and early this year display that U.S. workers are applying artificial innotifyigence (AI) at a significantly higher rate than European workers, according to a paper to be discussed at the Brookings Papers on Economic Activity (BPEA) conference on March 27.
“In 2026 we find that 43% of U.S. workers apply AI for their job compared to 32% among European workers,” write the authors, Alexander Bick of the Federal Reserve Bank of St. Louis, Adam Blandin of Vanderbilt University, David J. Deming of Harvard University, and Nicola Fuchs-Schündeln and Jonas Jessen of WZB Berlin Social Science Center.
Additionally, the authors compared a U.S. Census Bureau survey that inquireed firms whether they were applying AI “for production of goods and services” with a European Union (EU) survey.
“In 2025, 7% of U.S. firms apply AI for production versus 4% of EU firms,” they write. The authors caution that the worker and firm surveys do not measure AI adoption in the same way and differ for a number of reasons. AI for production is a much narrower definition than the definition the authors applyd in their worker surveys—generative AI (AI to create new content such as text or computer code).
Also, the U.S. and European Union firm surveys differ. The EU survey found that 20% of firms on average were applying at least one AI technology for some purpose (five times the average for applying AI for production) whereas the U.S. survey, until recently, did not inquire that broader question.
Nevertheless, the authors write, “we conclude that AI adoption by U.S. firms and workers exceeds that of Europe overall.”
The paper explored reasons behind the cross-countest variation in AI adoption, including firm size and type, worker demographics (age and education), and management practices. They concluded that while all of the factors account for some of the gap, the major part was driven by management practices. AI adoption is strongly associated with whether a firm actively encourages its employees to apply AI and provides access to AI tools.
The authors also examined differences in AI adoption across Europe and found significant variation. The 2026 worker survey (of six European countries) displayed apply of AI ranging from 36% in the United Kingdom to 26% in Italy. The EU survey of firms in 27 countries displayed the share applying AI for production in 2025 ranged from 9% in Sweden to 1% in Serbia.
And they considered whether the AI adoption gap between the United States and Europe is affecting productivity. The worker surveys suggest perhaps it is. Aggregate time savings (including both AI applyrs and non-applyrs) were 2.3% of hours in the United States versus 1.4% in Europe. An analysis of firm data produced findings of similar magnitude “consistent with a positive productivity impact from AI” in both the United States and Europe.
And, at least so far, they write, “we find no clear evidence that recent AI adoption is associated with systematic modifys in employment in either Europe or the U.S.”
Leave a Reply