After a period of modest productivity growth in the 2010s, U.S. labor productivity has increased since 2022, according to research from the Federal Reserve Bank of Kansas City. The rise in output per hour, a standard measure of labor productivity, is evident when comparing recent data with pre-pandemic trends. Chart 1 in the report shows that since late 2022, labor productivity has moved above its pre-pandemic trajectory and continued to climb through the second quarter of 2025. This period aligns with the commercial introduction of generative artificial intelligence (AI) tools.
Researchers analyzed whether this uptick signals a broader shift toward higher productivity and if it is linked to AI adoption across industries. They used industry-level data from the Federal Reserve Bank of Chicago and information on AI usage from the U.S. Census Bureau’s Business Trends and Outlook Survey.
The findings indicate that while overall productivity gains have accelerated, they are concentrated in a smaller set of industries rather than being widely shared. The four leading contributors during the so-called “gen-AI era”—retail trade, information, professional/scientific/technical services (PSTS), and real estate/rental/leasing—nearly doubled their average contributions compared to the previous decade. Some sectors such as mining and nondurable goods also shifted from minimal or negative contributions before the pandemic to positive ones more recently.
Within these broad sectors, specific sub-industries were responsible for most gains. For example, within information services, growth shifted toward data processing and hosting; within PSTS, computer systems design led improvements.
The distribution of productivity gains remains narrow. Chart 3 shows that only after including about 64 percent of value added does the cumulative contribution curve turn positive in the gen-AI era, indicating that a limited group of industries account for most net increases. Although less widespread than earlier periods, total productivity gains have more than doubled compared to pre-pandemic years.
To determine if AI adoption explains these changes, researchers compared industry-level AI use with productivity growth rates. While higher AI adoption correlates with faster growth within individual industries, it accounts for only a small portion of aggregate improvements. Not all high-adoption industries saw significant increases in their contribution to overall productivity; manufacturing’s contribution rose despite lower reported AI use.
The report concludes that although labor productivity has exceeded its pre-pandemic trend due to gains among select industries—and there is some alignment between AI adoption and industry growth—the aggregate impact remains limited as diffusion continues.
“The larger pickup and the positive AI adoption-growth link point to upside potential as adoption spreads, but the cross-industry pattern suggests diffusion is still incomplete,” write Nida Çakır Melek and Sydney Miller in their analysis.
Whether these recent gains signal a lasting change will depend on how broadly future advances spread across different sectors.
