Generative AI usage statistics: who uses AI at work, by role and country
Generative AI usage statistics from Gallup, the Fed, Eurostat, UK government and Pew, compared by role and country, with the realistic range to quote.
Klevere AI Team
AI Strategy Team
How many people use generative AI at work? The honest answer is somewhere between one in six and nearly one in two, depending on who you ask and how they ask. In the United States, Gallup found that 45% of employees used AI at work at least a few times a year by the third quarter of 2025, yet only 10% used it daily. Across the EU, Eurostat reported that 33% of people aged 16 to 74 had used generative AI tools in the three months before its 2025 survey, for any purpose.
This guide gathers the main published generative AI usage statistics for workers, compares them by country and by role, explains why they disagree, and gives a realistic range you can quote. It focuses on individual workers rather than companies. If you need the business-level picture, our separate guide to ai adoption statistics covers how many firms use AI. None of the figures below are Klevere data. Each comes from the publisher linked next to it.
Quick answer
About 40% to 45% of US workers use generative AI for work in some form, against roughly one in six working-age adults in the EU and one in three UK adults. Only about one in ten US workers uses it daily. Use is highest among younger, degree-educated staff in technology, finance and professional services.
The headline numbers side by side
Each line below gives the publisher, the figure, when the data was collected and what population it covers. Read the scope before you quote any of them, because a percentage without its denominator tells you very little.
Two patterns stand out straight away. Figures that count any use in a long window, such as a few times a year or the last three months, sit between 30% and 60%. Figures that count frequent or daily use sit between 9% and 23%. Both are true at the same time.
United States: broad but shallow use
The US has the richest run of worker-level data. Gallup has asked the same question since 2023, which makes its trend line the most useful. Its figures show use of AI at work a few times a year or more rising from 21% in the second quarter of 2023 to 40% in the second quarter of 2025 and 45% by the third quarter of 2025. Daily use rose from 4% to 10% over the same period.
The Federal Reserve's own monitoring lands in a similar place. A FEDS Note published in April 2026 put work-related generative AI use at about 41% of the workforce as of November 2025, while non-work use stood at about 50% of the population. Firm-level adoption is far lower, because one employee using a chatbot does not make a company an AI adopter.
Earlier research from economists at the St. Louis Fed found 28% of US workers using generative AI at work in August 2024. The researchers estimated average time savings of 5.4% of work hours for users. Pew, which asked specifically about chatbots such as ChatGPT, Gemini and Copilot in October 2024, found that 55% of workers rarely or never used them and 29% had not heard of them at all.
The gap between Pew's 16% and Gallup's 45% looks dramatic but is mostly explained by timing and definition. Pew measured chatbot use at least a few times a month in late 2024. Gallup measured any AI use at least a few times a year in mid 2025. A lower frequency threshold always produces a higher number.
Outside the US: UK, Europe and global surveys
United Kingdom and Europe
The UK government's largest public survey on technology found that 59% of UK adults had used a form of generative AI in the three months to early 2026. Usage was highest in London at 69% and lowest in North East England at 51%. Of the generative AI users in the online sample, 83% had used it in a personal setting and 54% for work purposes. That work figure is a share of users, not of all workers, so the share of the whole UK workforce using generative AI for work is lower, and the survey does not report it directly.
Eurostat's 2025 survey, the first EU-wide survey to ask individuals about AI, found that 33% of people aged 16 to 74 had used generative AI tools. The spread between member states is wide. Denmark led at 48%, followed by Estonia and Malta at 47% and Finland at 46%. The lowest rates were in Romania at 18%, Italy at 20% and Bulgaria at 23%. Outside the EU, Norway reached 56% and Switzerland 47%.
Eurostat also reports that in every main working-age group from 25 to 64, more than half of those using generative AI used it for work-related purposes. Put together with the 33% headline, that suggests somewhere in the region of one in six to one in five working-age Europeans use generative AI for work, with the Nordic and Baltic countries well above that and southern and eastern member states below it. Treat that as an estimate built from published figures rather than a statistic in its own right.
For businesses operating across borders, the country gap matters more than the average. Rollout plans, training and policies should reflect local readiness rather than assume one level everywhere.
Multi-country surveys
Microsoft's 2025 Work Trend Index is one of the few sources that covers many countries with the same questionnaire. It surveyed 31,000 workers across 31 countries between February and March 2025 and found that 45% of employees and 69% of leaders used AI regularly, defined as at least several times a week. It also found that 20% of employees and 29% of leaders said AI saved them at least an hour a day.
Two cautions apply. Microsoft sells AI tools, so its research is best used for direction rather than as the only figure you quote. Its sample is also full-time workers in each market, a narrower group than national statistics cover, which helps explain why its regular-use figure sits above Gallup's 23% for frequent use across all US employees.
Research from Denmark gives a useful independent check. Humlum and Vestergaard surveyed workers in 11 occupations judged highly exposed to AI and linked the answers to official register data. Their working paper reported that half of workers in those occupations had used ChatGPT, with younger, less experienced and higher-achieving workers leading. The peer-reviewed version in PNAS found that women were 16 percentage points less likely to have used the tool for work than men in the same occupations.
Usage by role and industry
Role and industry explain more of the variation than country does. Gallup's 2025 data shows the gap clearly: 76% of employees in technology or information systems used AI at work a few times a year or more, compared with 58% in finance and 57% in professional services. At the other end were retail at 33%, healthcare at 37% and manufacturing at 38%. Gallup describes the split as one between knowledge-based jobs and frontline positions.
The St. Louis Fed research measured intensity rather than reach. Computer and mathematics occupations used generative AI in nearly 12% of their work hours, compared with 1.3% for personal service occupations. By industry, information services had the largest share of hours assisted by generative AI at 14.0%, while leisure, accommodation and other services had the lowest at 2.3%.
Use is also spreading beyond the obvious desk jobs. A 2026 study published by the San Francisco Fed found at least one in five workers using generative AI in 80% of occupations, although adoption remained below 50% in most of them. The same study found that measures of how exposed a job is to AI explain only about half of the variation in actual use, which means two people in the same job title can have very different habits.
The pattern by role that comes through across sources is:
Leaders also use AI more than the people they manage. Microsoft found a 24 point gap between leaders and employees on regular use, and Pew found that only 12% of US workers said their employer encouraged the use of AI chatbots, with half saying their employer neither encouraged nor discouraged it.
Usage by age, education and gender
Age is the most consistent dividing line. In the EU, 64% of people aged 16 to 24 used generative AI tools in 2025, against only 7% of those aged 65 to 74. In the UK, DSIT found use most common among adults aged 16 to 44, falling steadily with age to 8% for those aged 85 or over. Pew found that 23% of US workers aged 18 to 29 used chatbots at least a few times a month, the highest of any age group.
Education follows close behind. Eurostat reports that 49% of people with a high level of formal education used generative AI tools, compared with 26% at a medium level and 22% with little or no formal education. Pew found US workers with a postgraduate degree were the most frequent chatbot users at 26%.
Gender gaps are smaller but persistent. Eurostat found men slightly ahead of women at 35% against 30%. The Danish study found a much larger gap of 16 percentage points within the same exposed occupations, which suggests that the difference is not only about who holds which job. For employers, this is a training and confidence question as much as a technology one.
Why the numbers disagree and the range to quote
If you have seen claims anywhere from 10% to 75%, the following four factors account for almost all of the difference.
A good rule is to quote one figure for reach and one for depth, always with the year and the population. For example: around four in ten US workers used generative AI for work by late 2025, but only about one in ten used it daily.
The realistic range to quote
Combining the sources above, a defensible summary for late 2025 to 2026 looks like this.
The direction of travel is clear in every series that has more than one wave. Use is rising, and the gap between occasional and daily users is where most of the remaining growth sits.
How we compiled this
We selected sources that publish their method, sample size and fieldwork dates, giving priority to national statistics agencies, central banks, government departments and peer-reviewed research, and using vendor research only alongside independent data. We opened each source to confirm the figures quoted. Where a source reported a share of users rather than a share of all workers, we have said so. The ranges in this guide combine published figures and are our interpretation, not new survey data.
What this means for your business
Many of your staff are probably already using generative AI, approved or not. Pew found that half of US workers said their employer neither encouraged nor discouraged chatbot use. That creates risk around confidential data and accuracy, and it also means the benefits are uneven and invisible. A clear ai usage policy is the cheapest first step.
The second takeaway is that occasional use does not produce the gains most leaders expect. The St. Louis Fed found that daily users reported far larger time savings than people who used the tools once a week. Moving people from occasional to habitual use usually takes role-specific workflows and training rather than another licence. If you plan to measure the payoff, our guide to ai roi explains how to set a baseline first.
How Klevere approaches it
When we work with a team, we start by finding out how people already use generative AI, which tools they use and where it saves time today. That baseline usually shows a handful of enthusiastic users, a large group of occasional users and some roles where nobody uses it at all. The useful work is turning individual habits into shared workflows and, where it makes sense, into custom AI agents that run a whole process rather than relying on each person to prompt a chatbot.
For teams that are earlier on, our ai consulting and training work focuses on the roles that the statistics show are lagging, with practical examples drawn from their own tasks. Across more than 50 projects in 12 industries and over 500 AI agents deployed, the pattern is consistent: adoption follows usefulness, and usefulness follows workflows designed around real tasks.
Frequently asked questions
How many employees use AI at work?
In the US, about 45% of employees said they used AI at work at least a few times a year in Gallup's August 2025 survey, and the Federal Reserve put work-related generative AI use at about 41% of the workforce in November 2025. Only around 10% use it daily. Multi-country surveys of knowledge workers report higher figures because they sample office staff only.
Which countries use generative AI the most?
In Europe, Norway leads at 56% of people aged 16 to 74, followed within the EU by Denmark at 48%, Estonia and Malta at 47% and Finland at 46%, according to Eurostat's 2025 data. The lowest EU rates were in Romania, Italy and Bulgaria. In the UK, 59% of adults used generative AI in the three months to early 2026.
Which jobs use generative AI the most?
Technology and information systems roles lead, with 76% of US employees in that sector using AI at work at least occasionally according to Gallup. Finance and professional services follow at close to 60%. Computer and mathematics occupations also spend the largest share of their hours using it. Retail, healthcare and manufacturing have the lowest use.
What percentage of workers use AI every day?
Around one in ten. Gallup found 10% of US employees used AI at work daily in August 2025, and the St. Louis Fed found 9% of US workers used generative AI every workday in 2024. Daily use is growing faster than occasional use but remains a minority habit in every published national survey.
Why do AI at work statistics vary so much?
Surveys define use differently, ask about different time windows and sample different groups. Any use in the past year gives a much higher figure than daily use, and surveys of full-time office workers report higher rates than national statistics covering all workers. Always check the year, the question and the population before comparing two numbers.
Are younger workers more likely to use generative AI?
Yes. Eurostat found 64% of 16 to 24 year olds in the EU used generative AI tools in 2025, against 7% of those aged 65 to 74. UK and US surveys show the same pattern, with use highest among adults under 45 and falling steadily with age. Higher education levels also strongly predict use.
If you want to know where generative AI could make the biggest difference across your own roles, and where your team's current use is creating risk, book a free AI audit. We will map how your people work today and show which processes are worth automating first.
Sources
- Gallup: AI Use at Work Rises
- Federal Reserve FEDS Notes: Monitoring AI Adoption in the U.S. Economy
- Federal Reserve Bank of St. Louis: The Impact of Generative AI on Work Productivity
- Federal Reserve Bank of San Francisco: What Work Does Generative AI Do?
- Pew Research Center: Workers' experience with AI chatbots in their jobs
- Microsoft Work Trend Index 2025: The year the Frontier Firm is born
- Eurostat Statistics Explained: Use of artificial intelligence by individuals
- UK DSIT: Public Engagement Survey 2025/2026
- Humlum and Vestergaard, PNAS: The unequal adoption of ChatGPT exacerbates existing inequalities among workers
- IZA Discussion Paper 16992: The Adoption of ChatGPT