US statistics office links worker AI use to stronger output growth, but cannot prove cause

The US Bureau of Economic Analysis published a spotlight article on 2 October 2026 about AI use and economic performance. The authors compared worker-reported AI use from Gallup Panel surveys with employer-reported adoption from the Census Bureau's Business Trends and Outlook Survey. The two sources agree closely at industry level, with a correlation of 0.93.

Worker-reported AI use rose from about 20 percent in mid-2023 to about 50 percent in early 2026. Frequent users rose from about 10 percent to over 25 percent. The authors then linked AI use by state and industry to data on employment, earnings and output. Regions and industries with higher AI use showed stronger real output growth after 2020. Employment changes were mostly positive but measured with less precision.

The authors say the evidence fits a link between AI and stronger output and productivity growth. They also say it is difficult to identify how much AI itself contributes.

This shows how hard it is to measure value from a new technology. Use is easy to count, but value appears later, unevenly, and mixed with other causes. Official statistics need several sources, such as workers and firms, to see where value is created. Even then, a correlation between use and output is not proof that the tool created the value.