US workplace AI use crossed a line in Q2 2026: 52% of employees now use AI in their role, up from 21% in Q2 2023, per Gallup’s May 6–20, 2026 survey of 22,573 employed adults. Firm-level counts read far lower, near 18%. This post lays out what each survey measures, how the numbers moved, and what workers report doing with the tools.
AI Adoption at Work
How Many US Workers Use AI at Work?
Gallup has tracked the same question quarterly since 2023, which makes it the cleanest time series available. Total use reached 52% in Q2 2026. Frequent use, defined as a few times a week or more, reached 30%. Daily use reached 15%.
| Quarter | Total AI users | Frequent users | Daily users |
|---|---|---|---|
| Q2 2023 | 21% | 11% | 4% |
| Q2 2024 | 27% | 12% | 4% |
| Q2 2025 | 40% | 19% | 8% |
| Q3 2025 | 45% | 23% | 10% |
| Q4 2025 | 46% | 26% | 12% |
| Q1 2026 | 50% | 28% | 13% |
| Q2 2026 | 52% | 30% | 15% |
Source: Gallup quarterly Gallup Panel workforce study, share of employed US adults, Q2 2023 through Q2 2026 (Q2 2026 fielded May 6–20, 2026, n=22,573)
Why AI Adoption at Work Estimates Range From 18% to 78%
Federal Reserve economist Jeffrey S. Allen compared three surveys in an April 2026 FEDS Note and found spreads that come down to who is counted. Firms and individuals are not the same denominator, and employment-weighted firm estimates run higher because large firms employ most people.
| Scope | Source | Date | Unit of analysis | Estimation type | Estimate |
|---|---|---|---|---|---|
| AI | BTOS | Dec. 2025 | Firms | Firm-weighted percentage | 18% |
| AI | SBU | Nov. 2025 | Firms | Employment-weighted percentage | 78% |
| GenAI/LLMs | RPS | Nov. 2025 | Individuals | Share of labor force | 41% |
| GenAI/LLMs | SBU | Nov. 2025 | Firms | Employment-weighted percentage | 54% |
Source: Federal Reserve Board, “Monitoring AI Adoption in the U.S. Economy,” FEDS Notes, April 3, 2026, citing Census Bureau BTOS, Real-Time Population Survey, and Atlanta Fed Survey of Business Uncertainty
Wording shifts the answer too. Pew Research Center asked how much of a person’s work is done with AI and got 21% saying at least some, up from 16% a year earlier, in a September 2–8, 2025 survey of 5,010 US workers. Another 65% said they do not use AI much or at all. The Census Bureau widened its own question in November 2025 to cover AI use in any business function, replacing a narrower one about producing goods or services.
Source: Pew Research Center, “About 1 in 5 U.S. workers now use AI in their job,” Sept. 2–8, 2025 survey; Federal Reserve Board FEDS Notes, April 3, 2026
AI Adoption at Work by Organization
Organizational integration lagged individual use, then jumped. Employees saying their organization has integrated AI tools went from 41% in Q1 2026 to 47% in Q2 2026, the sharpest quarterly move Gallup has recorded. One in five still do not know either way, and that share has barely budged.
SHRM lands on the same 47% figure from a different sample. In its March–April 2026 survey of 5,875 US workers, 47% said their organization has implemented AI, and an equal 47% said it has not.
Worker-Level AI Use Split
SHRM’s 41% work-use figure breaks into 8% who use AI for work only and 33% who use it for both work and personal tasks. A quarter use AI for personal purposes only, and 34% use no AI tools at all. Small-company workers, who often run a compact home office setup, sit in organizations least likely to have formal AI rollouts.
Source: SHRM, “Navigating AI in the Workplace: 2026,” survey of 5,875 US workers, March–April 2026
What Workers Use AI For at Work
Language tasks dominate. Writing and editing tops Gallup’s list at 51% of AI users, followed by search or research at 49%. Technical applications sit far lower. Respondents could select more than one application, so these shares overlap and cannot be added together.
| Application | Share of AI users |
|---|---|
| Writing and editing | 51% |
| Search or research | 49% |
| General assistance or problem solving | 39% |
| Knowledge or information management | 31% |
| Email or communication management | 29% |
| Meeting assistance or transcription | 19% |
| Data science or analytics | 18% |
| Presentation or slide deck creation | 17% |
| Image, video, or audio generation | 17% |
| Coding assistance | 16% |
| Task, scheduling, or project management | 16% |
| Automation or process automation | 16% |
| Other | 3% |
Source: Gallup, Q2 2026 workforce study, AI users among employed US adults, fielded May 6–20, 2026
Frequent users pull ahead most on specialized tasks. They are nearly three times as likely as infrequent users to use AI for coding assistance, 22% against 8%, the gap most visible at a developer workstation.
Does AI Adoption at Work Improve Productivity?
Gallup found reported productivity gains scale with how many different tasks a worker applies AI to. Among people using it for one or two purposes, 45% reported a positive effect on productivity. Among those using it for seven or more, 90% did. Gallup cautions this correlation does not prove that adding use cases causes the gain.
By application, the highest ratings came from coding assistance and automation, each at 77%, ahead of writing and editing at 68% and search or research at 65%.
Source: Gallup, Q2 2026 workforce study, AI users among employed US adults, fielded May 6–20, 2026
Hours Saved Against Hours Spent Fixing Output
SHRM asked workers to estimate both sides of the ledger. The average AI user reported saving about six hours a week and spending about four hours a week fixing AI output. Seniority widens both figures.
| Job level | Hours saved per week | Hours correcting AI output |
|---|---|---|
| Individual contributors | 4 | 3 |
| Managers | 7 | Not published |
| Directors and above | 9 | 6 |
| All AI users | 6 | 4 |
Source: SHRM, “Navigating AI in the Workplace: 2026,” survey of 5,875 US workers, March–April 2026 (self-reported estimates)
SHRM suggests redirecting saved time toward complex problem-solving, which is easier in a space built for long stretches of deep work and free of the clutter that breaks concentration. Half of workers said AI raised performance expectations, and 54% said the pace of work increased.
AI Adoption at Work by Industry
The Real-Time Population Survey, cited in the Fed note, measures work-related generative AI use among individuals by sector as of November 2025. Information leads at 70%. Accommodation and food services trails at 21%.
Source: Real-Time Population Survey (Bick, Blandin, and Deming), November 2025, cited in Federal Reserve Board FEDS Notes, April 3, 2026
Firm-level BTOS data ranks these sectors similarly but far lower, with professional services near 33% and financial near 30% of firms. Allen reads early AI use as skewing toward cognitive and analytical work, the kind done at a two-screen desk rather than on a shop floor.
AI Policies, Shadow AI, and Rule-Breaking
Among organizations that have adopted AI, fewer than half have written rules for using it. SHRM found 47% have policies regulating AI use, rising to 56% at large organizations and falling to 36% at small ones.
Thirty percent of workers said they knowingly broke their organization’s AI rules, usually by using tools still under review. Among them, 31% would do it again to save time, and 46% of all workers said policies block them from testing new tools. Staff working from a personal home office face fewer checks on which account they open.
Source: SHRM, “Navigating AI in the Workplace: 2026,” survey of 5,875 US workers, March–April 2026
FAQ
How many US workers use AI at work in 2026?
Gallup’s Q2 2026 survey of 22,573 employed US adults found 52% use AI in their role at least a few times a year. Thirty percent use it a few times a week or more, and 15% use it daily.
Why do AI adoption at work statistics disagree so much?
They count different things. The Census BTOS measures firms and reported 18% at year-end 2025. The Atlanta Fed’s employment-weighted estimate reached 78%, because large firms employ most workers. Individual surveys land between.
Do Reddit discussions about AI at work match the survey numbers?
Reddit threads are self-selected, not measured samples. Gallup’s probability-based panel put daily AI use at 15% and total use at 52% in Q2 2026, well short of the near-universal adoption forum discussion often implies.
Reddit posts describe hiding AI use from managers. How common is that?
SHRM’s March–April 2026 survey found 30% of workers knowingly broke their organization’s AI rules, usually by using tools still under review. Among those, 31% said they would break the policy again to save time.
How much time does AI actually save workers?
SHRM workers self-reported saving about six hours a week and spending about four hours correcting AI output. Directors and above reported nine hours saved and six spent fixing; individual contributors reported four and three.
Sources
https://www.gallup.com/workplace/712736/organizational-adoption-jumps-six-points.aspx
https://www.shrm.org/topics-tools/research/navigating-ai-in-the-workplace/full-report
