Distinguish 52% overall workplace AI use from the 15% daily habit

Among United States employees, daily use of artificial intelligence at work is still a minority habit inside a much larger group that has merely encountered the tools during the year rather than built them into the working day. Gallup’s May 2026 indicator put daily use at 15% of employees, use several times a week or more at 30%, and use at least a few times a year at 52%. The spread inside a single survey is the finding. A yearly acquaintance is not a daily job.

Those three May 2026 shares come from one Gallup indicator and function as nested frequency thresholds rather than as three separate populations of employees who never overlap. Anyone who uses the tools every day is already inside the group that uses them several times a week or more, and that larger group sits inside the still larger group that has used them at least a few times during the year, so the May 2026 indicator is a set of nested bands rather than three rival headcounts. The gap that matters is the drop from a yearly threshold that covers just over half of employees to a much smaller daily core. Industry talk often treats any workplace contact with the tools as if it were the same fact as daily use, and Gallup’s own bands show that those are different facts. The yearly band is the one that can be mistaken for saturation. More than half of employees, on Gallup’s May 2026 indicator, have used the tools at least a few times that year, which is already a majority on a yearly count. Frequency is the stricter test of the three Gallup bands on that indicator. Several times a week or more is a smaller group, and every day is smaller again.

Late 2025 on the same tracker

Gallup had already published a workplace reading on January 25, 2026 covering how often employees used the tools in the fourth quarter of 2025, when daily use stood at 12% of employees and frequent use stood at 26%. Daily use is the comparable item across the two Gallup releases: it was 12% at the end of 2025 and 15% in the May 2026 indicator. The January release labels its broader band as frequent use rather than as use several times a week or more, so the 26% figure and the later 30% figure keep their own labels. By May 2026, daily use among United States employees was still a minority of the people Gallup counted, and it had been a minority in the fourth quarter of 2025 as well.

Pew counted any job use

Pew Research Center, in a short read dated October 6, 2025, measured whether United States workers perform at least part of their job with artificial intelligence, and that survey put the share at 21%, which Pew also stated as 1 worker in 5, up from 16% in the previous Pew reading named in the same publication. The 16% figure arrives without its own field month. Pew is counting any job use and Gallup is counting how often people use the tools, so the percentages are answers to different questions rather than two readings of one rate.

Finance moved on the academic panel

Frequency can also move sharply inside one sector of an academic panel while the overall reading for that same panel stays lower, which is the pattern the 2026 conference paper shows for finance against the panel as a whole. A 2026 National Bureau of Economic Research conference paper, listed as f248303 and issued without a public URL, reported frequent use in finance at 8.4% in 2023, a year the paper does not break into months, and at 44.9% in the first quarter of 2026. The same paper put the panel’s overall figure at 27.4%, without naming a month for that overall reading. That overall rate is not a finance rate. The finance movement is evidence about one industry on one academic panel, not a warrant to treat every workplace as if it had made the same jump in frequent use during those years. The 2023 finance reading is also more than two years old relative to the 2026 paper that reports it, and nothing in that paper converts the finance jump into a picture of daily use across the whole United States workforce.

What employee surveys can say

These figures describe the United States and record what employees say about their own work, not what workplace systems log. They do not say whether a firm has adopted the tools as a company, and they cannot be read as a substitute for that firm-level count. Firm-level surveys of whether a business uses artificial intelligence at all answer that separate question, and they are laid out in AI at work: adoption by the numbers. On the employee side, Gallup’s May 2026 indicator separates these frequencies: 52% of employees used the tools at least a few times a year, 30% used them several times a week or more, and 15% used them every day. The survey bands measure any workplace contact and daily use as distinct facts.