“Half of workers now use AI” and “one in five workers use AI” can both appear in the same month, and both can come from respectable surveys. The difference usually lies in who was asked, what the question said and when the survey ran. For anyone planning to roll AI out at work, or trying to benchmark their own team, that is worth understanding before quoting a number.
This article summarises what Gallup and Pew Research Center surveys say about AI at work, with sample sizes, fieldwork dates and caveats. All of the figures concern US adults or employees, and we read some of them through news coverage, not the original reports. This is a news summary and a reading guide, not a forecast.
Key takeaways
- Gallup reported that 52 percent of US employees used AI in their role in a survey fielded 6 to 20 May 2026, with 30 percent using it frequently and 15 percent daily.
- Gallup’s earlier survey, fielded 4 to 19 February 2026, put use at 50 percent, so the trend over the year was upward but the two quarters are close.
- Pew reported in June 2026 that 21 percent of US workers use AI in their jobs, up from 16 percent in 2024, while 49 percent of all US adults use AI chatbots.
- The headline numbers differ because the surveys ask different questions of different groups, so do not compare them as if they measured the same thing.
- Positive productivity ratings in Gallup’s data rose with the number of different ways people used AI, but this is self-reported and shows association, not proof of cause.
Summary
Across the surveys we found, AI use at work is common and rising, but the picture depends on the question. Gallup, asking US employees whether they use AI in their role in some way, found about half saying yes by early 2026 and a little over half by May. Pew, asking workers whether they use AI in their jobs, found about one in five. Both found growth. Both also found that frequent, daily use is much rarer than occasional use, and that organisations often lag individuals.
The second theme is a gap between use and transformation. In Gallup’s first-quarter data, only about one in ten employees at AI-adopting organisations strongly agreed that AI had transformed how work gets done. Adoption is wide but often shallow.
Surveys at a glance
| Survey | Fieldwork | Sample | Published | Headline finding |
|---|---|---|---|---|
| Pew, US workers | 7 to 13 October 2024 | 5,273 employed US adults | 25 February 2025 | 16 percent say at least some of their work is done with AI |
| Gallup, Q1 2026 | 4 to 19 February 2026 | 23,717 US employees | Reported 15 April 2026 | 50 percent use AI at work in some capacity |
| Pew, US adults | February 2026 | 5,119 US adults | Reported 17 June 2026 | 49 percent of adults use AI chatbots; 21 percent of workers use AI in jobs |
| Gallup, Q2 2026 | 6 to 20 May 2026 | 22,573 employed US adults | 20 July 2026 | 52 percent use AI in their role; 47 percent say their organisation has integrated AI |
Gallup, first quarter 2026
TechRadar reported on 15 April that Gallup’s survey, fielded from 4 to 19 February 2026 with 23,717 US employees, found 50 percent using AI tools at work in some capacity. The article describes this as a 4 percent rise from the previous quarter and a 21 percent rise from three years earlier. It also reports:
- 13 percent use AI daily and 28 percent use it a few times a week or more.
- 41 percent say their organisation has integrated AI to improve its practices, up three points from the previous quarter.
- About one in ten employees at AI-adopting organisations strongly agree that AI has transformed how work gets done in their company.
- 27 percent of employees at AI-adopting firms report large or very large disruptive changes to their workplace in the past year, compared with 17 percent at non-AI firms.
- 18 percent of all employees say it is very or somewhat likely their job will be eliminated within five years due to AI or automation.
- Gallup reports that most AI users see productivity and efficiency gains, though often at the task level rather than across the organisation.
Gallup, second quarter 2026
Gallup’s own article, published on 20 July by Andy Kemp, reports a survey fielded 6 to 20 May 2026 with 22,573 employed US adults from the Gallup Panel, with a margin of error of plus or minus 0.9 percentage points. Findings include:
- 47 percent of employees say their organisation has integrated AI tools, up from 41 percent in Q1. 33 percent say it has not and 20 percent do not know.
- 52 percent use AI in their role, 30 percent use it frequently (a few times a week or more) and 15 percent use it daily.
- Among users, the most common uses are writing and editing (51 percent), search or research (49 percent) and general assistance or problem-solving (39 percent).
- Coding assistance and automation each reach 16 percent of users. Frequent users are nearly three times as likely to use AI for coding assistance (22 percent against 8 percent) and automation (21 percent against 8 percent).
| Measure (Gallup) | Q1 2026 (Feb) | Q2 2026 (May) |
|---|---|---|
| Use AI at work in some way | 50 percent | 52 percent |
| Use a few times a week or more | 28 percent | 30 percent |
| Use daily | 13 percent | 15 percent |
| Organisation has integrated AI | 41 percent | 47 percent |
The Q1 figures come from TechRadar’s report and the Q2 figures from Gallup’s own article. The question wording may not be identical between the two, so read the change as indicative.
Pew Research Center
US adults and chatbots (February 2026)
The Next Web reported on 17 June on a Pew Research Center report based on a February 2026 survey of 5,119 US adults. It says 49 percent of adults use AI chatbots, up from about a third in 2024, and that 21 percent of workers use AI in their jobs, up from 16 percent in 2024. Other figures reported: 71 percent expect AI to make personal data less secure, about two thirds say AI is advancing too quickly, and 51 percent of adults do not use chatbots, with 60 percent of those citing disinterest. Among chatbot users the top uses were information search at 42 percent, entertainment at 25 percent and image or video creation or editing at 24 percent. The article also reports usage shares among chatbot users by product, which we leave out because market share data from a single survey is a thin basis for conclusions.
US workers (October 2024)
Pew’s own February 2025 report on a survey of 5,273 employed US adults, fielded 7 to 13 October 2024, found that 16 percent said at least some of their work is done with AI, and another 25 percent said they did not use it much now but some of their work could be done with it. 52 percent were worried about the future impact of AI in the workplace and 36 percent hopeful. 32 percent expected fewer job opportunities for them in the long run. About one in ten workers used chatbots every day or a few times a week. Among chatbot users, 57 percent used them for research, 52 percent for editing written content and 47 percent for drafting. This is the baseline against which the 2026 figure of 21 percent can be read.
Why the numbers differ
Gallup’s 50 to 52 percent and Pew’s 21 percent describe different things, and we cannot fully reconcile them from the sources we opened. Likely reasons, which are our interpretation and not stated by the pollsters:
- Wording. “Use AI tools at work in some capacity” is broader than “some of my work is done with AI.”
- Population. Gallup surveyed employees from its panel. Pew surveyed all US adults and then reported on workers.
- Timing. Pew’s 21 percent comes from a February 2026 survey, while Gallup’s latest is from May.
- Method. Panels, weighting and mode differ.
| Question to ask of any statistic | Why it matters |
|---|---|
| Who was surveyed? | US employees, all adults or one industry give different answers |
| What exactly was asked? | “Ever use” and “daily use” are very different measures |
| When was it fielded? | Adoption figures date quickly |
| How large is the sample? | Small samples carry wider margins of error |
| Is it self-reported? | People may over or under-state their use |
What the surveys say about productivity
Gallup’s second-quarter article reports positive productivity ratings from users: 77 percent for coding assistance and automation, 76 percent for slide deck creation, 75 percent for data science or analytics, 68 percent for writing and editing and 65 percent for search or research. It also reports that positive ratings rise with the breadth of use: 45 percent among people with one or two use cases, 66 percent for three or four, 78 percent for five or six and 90 percent for seven or more.
These are self-reported views among users, not measured output. People who find AI useful may try it in more ways, so the pattern could run in either direction. Gallup’s first-quarter report also stresses that gains often occur at the task level, not across the organisation. If you want evidence for your own team, how to measure whether an AI tool saves time describes a simple way to gather it.
Worries and job security
The surveys also record unease. Gallup’s Q1 figure was 18 percent of employees saying it is very or somewhat likely their job will be eliminated within five years because of AI or automation. Pew’s June 2026 coverage reports that about two thirds of adults say AI is advancing too quickly and 71 percent expect personal data to become less secure. Pew’s 2024 worker survey found more workers worried than hopeful. None of these figures predicts what will happen. They show how people feel, which matters for rollout.
What it means for tool buyers and team leads
- Your team is probably already using AI. With around half of employees reporting use, assume some of your staff use tools you have not approved. A clear policy helps; see how to write an AI use policy for your team.
- Organisational adoption lags individual use. The gap between 52 percent individual use and 47 percent saying their organisation has integrated AI is small in these data, but only about one in ten at adopting firms strongly agreed work had been transformed in Q1.
- Training matters. Gallup found positive ratings rise with the variety of use cases, which suggests that learning more ways to use a tool may go with better perceived results. See how to train your team to use AI well.
- Do not copy a benchmark blindly. A national statistic is a weak guide to a small team in a specific industry.
- Pilot before you scale. A structured trial, as in how to run an AI tool pilot in 30 days, gives you your own numbers.
How to benchmark your own team
| Step | Action |
|---|---|
| 1 | Ask a short anonymous question set: which tools, how often, for what tasks |
| 2 | Use the same wording each quarter so you can compare over time |
| 3 | Separate “ever used” from “weekly” and “daily” use |
| 4 | Ask what data goes into the tools, to spot risk |
| 5 | Compare self-reported benefits with a measured task, such as time to draft a standard document |
| 6 | Share the results with the team and act on them |
If you are still choosing tools, how to choose an AI tool sets out a selection method.
What to watch next
- Gallup’s next quarterly update, to see whether the rise from 41 to 47 percent in organisational integration continues.
- Whether daily use, at 15 percent in May, moves faster than occasional use.
- Surveys outside the United States, which we did not find in this research.
- Measured productivity studies, as distinct from self-reported ratings.
These are questions to watch, not predictions.
Frequently asked questions
How many workers use AI?
It depends on the survey. Gallup reported 52 percent of US employees using AI in their role in May 2026, and Pew reported 21 percent of US workers using AI in their jobs from a February 2026 survey. The questions differ.
Why do Gallup and Pew differ so much?
Probably because of different wording, populations, timing and methods. We cannot confirm the exact cause from the sources we read.
Do these figures apply outside the US?
The surveys we found cover US adults or employees. We did not find comparable international data in this research.
Does more AI use mean more productivity?
The data show positive self-reported ratings rising with the number of use cases, but that does not prove cause. Measure your own results.
Should I worry about job losses?
Gallup reported 18 percent of employees thinking job elimination within five years was very or somewhat likely. That is a measure of perception, not a forecast.
Conclusion
The 2026 survey evidence points to widespread but uneven use of AI at work, with about half of US employees reporting some use in Gallup’s data and a smaller share in Pew’s differently worded question. Daily use and deep organisational change are much less common. Use the numbers carefully, ask your own team, and start with a clear policy and a small pilot. A sensible first read is how to write an AI use policy for your team.
Sources
- Gallup, Organizational AI Adoption Jumps Six Points, 20 July 2026, accessed 11 November 2026
- TechRadar Pro, Is this the tipping point for AI at work, 15 April 2026, accessed 11 November 2026
- The Next Web, Pew Research on Americans, AI chatbots and distrust, 17 June 2026, accessed 11 November 2026
- Pew Research Center, U.S. workers are more worried than hopeful about future AI use in the workplace, 25 February 2025, accessed 11 November 2026
