AI at Work Statistics 2026: 39% of US Workers Used AI Last Week, and Three Surveys Put Daily Use Near 13%

Red figures: our estimates from the Census Bureau’s Household Trends and Outlook Pulse Survey public-use file, March 13 to 30, 2026, 6,277 workers. 95% margins: ±4.1, ±3.7 and ±2.5 points.

Same quarter, four surveys, three questions

Every major survey of AI at work fielded in the first quarter of 2026 is shown below, grouped by how often the question asks about. Within each group the figures come from different questions and different samples, so they are shown side by side and never averaged.

The pattern is consistent. On daily use the three surveys that ask about it agree within 1.0 point: 13.3% (Census), 13% (Gallup) and 12.2% (Real-Time Population Survey). On weekly use they spread over 10.6 points, and on “ever” or “any” use over 12.7 points. The looser the question, the more the answers diverge.

Is AI use at work rising?

Yes, on every series that has more than one reading, though the newest data point is a reminder that quarterly surveys wobble. Gallup’s share of employees using AI at least a few times a year went from 21% in May 2023 to 52% in May 2026, and its weekly-or-more share from 11% to 30%. Its daily share went from 4% to 15%. Comparing May with May to avoid seasonal differences in fieldwork, Gallup’s total rose 12 points in a year and the Real-Time Population Survey’s work-use rate rose 10.1 points, from 35.1% to 45.2%.

The Real-Time Population Survey’s August 2026 reading then fell to 40.5%, 4.7 points below its May 2026 peak, with weekly use down from 39.2% to 34.8%. That is a single quarter: the same series dipped from 33.3% to 31.0% in late 2024 before climbing for six straight readings, and the August 2026 figure is still 3.2 points above August 2025. Gallup has not yet published a third-quarter 2026 reading to confirm or contradict it. Pew Research Center, which asks how much of a worker’s job is done with AI, found the share answering “some” or more rose from 16% in October 2024 to 21% in September 2025.

The close agreement on daily use is recent. In May 2025, Gallup’s daily share was 8% against the Real-Time Population Survey’s 11.2%. Gallup’s has nearly doubled since, while the RPS daily rate has stayed between 9.1% and 13.7% in every survey since August 2024, which is why the two lines cross around the start of 2026.

Lines join published readings; points are single surveys. Census whiskers are 95% confidence intervals. RPS readings are plotted at the survey month (the middle month of each FRED quarter). All panels use the same 0 to 60% scale.

Every survey, its wording and its latest reading

Which occupations use AI at work

Occupation is the sharpest dividing line in the Census data. In computer occupations, 80.4% of workers used AI at work in the past week and 49.2% used it every day. In security, food, cleaning, personal care the weekly share was 9.2%, and in construction, trades, production, transportation it was 10.8%: a gap of roughly 9 to 1 between the top and bottom groups.

± values are 95% margins of error. Groups follow the Census occupation question, which combines related major occupation groups.

Who uses AI at work: education, age and telework

Education tracks AI use closely. Workers with a graduate degree reported weekly use at 60.4%, against 20.0% for those with a high school diploma or less. Gallup’s job quality study found the same split on its own question (college graduates 40% weekly or more versus 17% without a degree), and Pew found its “some or more” share rising fastest among workers with a bachelor’s degree (20% to 28%).

Telework is the second big divide. Among workers who did not work from home at all in the past week, 26.7% used AI at work that week; among those who teleworked one to two days it was 68.3%, and among full-time remote workers 61.5%, with 30.3% using it every day. That matches Gallup’s finding that total AI use reached 66% in remote-capable roles against 32% in roles that cannot be done remotely. Only 31% of workers in the Census sample teleworked at all that week, so most of the workforce sits in the lower-use group; our remote work statistics track how many people work from home, by job and state, in federal survey data.

Age runs against expectations. In the Census data, workers aged 35 to 49 were the most likely to have used AI at work (65.3%), ahead of workers under 35 (50.3%), a statistically significant gap; usage then falls to 40.1% among workers 65 and older. Gallup’s job quality study likewise found younger workers are not among the most frequent users. Men and women used AI at work in the past week at similar rates (40.2% and 37.0%, not a significant difference), but men were more likely to use it every day (17.0% against 9.5%).

What workers use AI for

The most common uses are general office work, not specialist automation. The top three tasks are searching for information or technical help (36.9% of workers), writing communications, documentation, or instructions (32.5%) and generating ideas (32.2%). Coding or modeling, at 11.0%, ranks below customer support.

Census HTOPS, March 2026. Each task’s share uses workers who answered that item. Workers could select any number of tasks.

How much time AI saves at work

Workers who used AI at work in the past week were asked how many extra hours they would have needed without it. Of those, 61.4% said at least one hour, 31.4% said one to two hours and 22.1% said four hours or more, while 13.4% said it saved no time or made the work take longer. These are self-reports, so they measure perceived savings rather than measured output.

The Real-Time Population Survey turns similar questions into a share of all work hours, including people who do not use AI. Its estimate of hours spent working with generative AI peaked at 6.3% in May 2026 and stood at 5.3% in August 2026; the hours saved came to 2.1% of all work hours, up from 1.4% in November 2024.

Source: Bick, Blandin and Deming, Real-Time Population Survey, via FRED. Assisted hours are the midpoint of bracketed responses.

AI use by industry

Gallup publishes its own industry series. Technology leads on every tier: 85% of technology employees used AI a few times a year or more in May 2026 and 42% used it daily, against 40% and 12% in retail. Gallup’s industry groups do not match the Census occupation groups, so the two tables should not be compared row by row.

Source: Gallup, AI use trends by industry (published workbook). Sorted by frequent use in May 2026.

Workers vs businesses: why the business figure is lower

The Census Bureau also asks businesses about AI. In the two Business Trends and Outlook Survey collections that overlap the worker survey’s March field period, 18.9% and 19.1% of employer businesses said they had used AI in a business function in the previous two weeks. By October 4, 2026 it was 23.8%. Our small business AI adoption report breaks that series down by firm size, sector and state.

The business and worker figures cannot be subtracted from each other. The business survey counts firms, so a four-person shop weighs the same as a company with 10,000 employees. The worker surveys count people. Workers also use AI that their employer may not know about: in Gallup’s panel, 23% of employees said they did not know whether their organization had implemented AI in the third quarter of 2025, and one in five remained unsure in May 2026, when 47% said their organization had integrated AI tools.

What this means for MSPs

• Plan for the weekly cohort, not the whole headcount. About 39% of workers use AI in a typical week, but the rate is 80% in computer occupations and 63% in management and business roles, against roughly 1 in 10 in frontline trades and services. Size paid AI assistant pilots and support capacity to the desk-based, remote and hybrid staff who already use these tools.
• Assume unmanaged use. Gallup finds a fifth of employees do not know whether their employer has adopted AI, while half use it. Discovery of which AI tools staff already use, followed by data-handling rules, is a concrete first engagement.
• Write data rules around search, writing and summarizing. Searching for information (36.9% of workers), writing communications (32.5%) and summarizing or translating (31.1%) are three of the four most common Census tasks, so acceptable-use and data-handling rules should cover them first.
• Quote the right number. Figures as high as 89% of workers circulate in search results for this topic; check the wording before repeating one. The March 2026 Census figures are about 13% daily, 39% weekly and 56% ever.
• Watch the workforce side as well as the tooling. 18% of employees told Gallup in early 2026 that their job was likely to be eliminated by AI or automation within five years. Our AI layoffs tracker follows the job cuts employers have attributed to AI so far.

Frequently asked questions

What percentage of people are using AI at work?

In the Census Bureau’s Household Trends and Outlook Pulse Survey (March 13 to 30, 2026), 55.6% of US workers had used AI for at least one of 11 work tasks, 38.6% used AI at work on at least one day in the past week, and 13.3% used it every day. Gallup’s May 2026 panel found 52% use AI in their role a few times a year or more, 30% a few times a week or more and 15% daily.

Is AI use at work increasing?

Yes, in both long-running series. Gallup’s total-use share rose from 21% (May 2023) to 52% (May 2026). The Real-Time Population Survey’s work-use rate rose from 33.3% in August 2024 to a peak of 45.2% in May 2026, then read 40.5% in August 2026, still 3.2 points above August 2025.

Which jobs use AI the most?

In the Census March 2026 data, 80.4% of workers in computer occupations used AI at work in the past week, followed by management, business, financial and legal occupations at 63.0%. The lowest rates were in security, food, cleaning and personal care (9.2%) and construction, trades, production and transportation (10.8%).

How many workers use AI every day?

About 1 in 8. In the first quarter of 2026, the Census survey put daily use at 13.3% of workers, Gallup at 13% and the Real-Time Population Survey at 12.2% (every workday last week). Gallup’s May 2026 reading was 15%.

How much time does AI save at work?

Among workers who used AI at work in the past week, 61.4% told the Census Bureau it saved them at least an hour of work that week and 22.1% said four hours or more; 13.4% said it saved no time or added time. Across all work hours, the Real-Time Population Survey estimates time savings of 2.1% of hours in August 2026.

Why do AI at work surveys give such different numbers?

They ask different questions of different people. ‘Ever used’ questions produce the highest figures (55.6% in Census data), weekly questions land between 28% and 38.6%, and daily questions agree closely near 13%. The Census asks about AI in general, the Real-Time Population Survey only about generative AI, and Gallup only surveys people who work for organizations.

Methodology

References

1. U.S. Census Bureau, Household Pulse Survey / HTOPS public-use files (HTOPS HPS 2603). Accessed October 2026.
2. U.S. Census Bureau, HTOPS Household Pulse User Notes, March 2026 (revised weighting, September 10, 2026). Accessed October 2026.
3. U.S. Census Bureau, “Does Using Artificial Intelligence Save Time at Work?” (August 11, 2026, revised September 24, 2026). Accessed October 2026.
4. U.S. Census Bureau, Business Trends and Outlook Survey. Accessed October 2026.
5. Bick, Blandin and Deming, Generative Artificial Intelligence, Adoption Rate for Work: Employed Adults (FRED), with the companion series Use Last Week for Work, Daily Use for Work, Work Hours Assisted and Time Savings. Accessed October 2026.
6. Bick, A., Blandin, A. and Deming, D. (2026), “The Rapid Adoption of Generative AI,” Management Science, and the GenAI Adoption Tracker. Accessed October 2026.
7. FRED Blog, “Does generative AI save time at work?” (August 27, 2026). Accessed October 2026.
8. Board of Governors of the Federal Reserve System, FEDS Notes, “Monitoring AI Adoption in the U.S. Economy” (April 3, 2026). Accessed October 2026.
9. Gallup, Indicator: Artificial Intelligence, with its AI use trends by industry workbook. Accessed October 2026.
10. Gallup workforce articles: June 15, 2025; December 14, 2025; January 25, 2026; April 12, 2026; July 20, 2026. Accessed October 2026.
11. Gallup, “AI Benefits at Work Unevenly Distributed” (American Job Quality Study, October 6, 2026). Accessed October 2026.
12. Pew Research Center, “About 1 in 5 U.S. workers now use AI in their job, up since last year” (October 6, 2025), and its topline. Accessed October 2026.