
ChatGPT
AI and Time at Work: For a Few Workers, Most of the Hours Saved
How many workers say AI saves them time, how much it saves, and who gets the hours
David Lazer, Ata Aydin Uslu, John Wihbey, Alisa Zomer, Sarah H. L. Bernt, Jack X. Howard, Claire Z. Li, Cassidy Waldrip, Elijah Bader-Gregory
Research Report
|
November 18, 2026
Executive Summary
Key Findings
This report draws on the Northeastern University AI & Society Survey of 38,375 adults, fielded in the summer of 2026 in all 50 states and the District of Columbia (see methods statement). As the use of AI tools spreads through the workplace, there is much to understand about who uses them, who receives formal training, and whether they deliver meaningful gains. This report examines how workers across the United States report using AI and how much time they say it saves them.
The central finding is that AI saves time for American workers, but the hours saved are very unevenly spread. Most American workers (55%) do not use AI regularly. A third of workers (33%) use AI at least monthly and say it saves them at least one hour a week. A small share of workers (9%) say AI saves them eight or more hours a week, and they account for more than half of all the hours saved. Time savings rise steeply with education, income and seniority, and they vary widely across industries. Men are more likely than women to report saving time, and workers ages 25 to 44 more likely than younger and older workers. These results raise questions about whether AI tools will amplify existing inequalities in the U.S. economy and how worker retraining could play a role as new tools are deployed.
Who counts as an AI user in this report. Workers are respondents who are employed full- or part-time,
self-employed, or doing gig or contract work. “AI users” are workers who use AI for their work at least once a
month, who make up 45% of workers, or 9,065 of the survey’s respondents. The 10% of workers who report having
tried AI only once or twice are not counted as users. A worker who “saves time” is an AI user who says AI saves
them at least one hour a week. Every figure describing a share of AI users is based on the monthly-or-more group;
figures for all workers include all 20,275 respondents.
-
A third of workers say AI saves them time, but most of the hours go to a small share of workers. Overall, 33% of workers use AI at least monthly for work and say it saves them an hour a week or more. Half of all the hours saved go to 8% of workers. When each hour is valued by household income, the 10% of workers whose saved time is worth the most account for 68% of the total.
-
Workers with more education and higher incomes are the most likely to save time with AI. Among all workers, 59% of those with a graduate degree and 54% of those in households earning $100,000 or more say AI saves them at least an hour a week, compared with 20% of high school graduates and 19% of those in households earning under $50,000. Part of the gap reflects use: 74% of workers with a graduate degree use AI for work at least once a month, compared with 30% of high school graduates.
-
Men are more likely than women to save time, and workers ages 25 to 44 are more likely than younger or older workers. Some 38% of working men and 27% of working women say AI saves them at least an hour a week. Among workers ages 25 to 44, 39% report saving time, compared with 28% of workers ages 45 to 64 and 25% of workers ages 18 to 24.
-
The youngest and least senior workers are also the most likely to see AI as a threat. Some 27% of workers ages 18 to 24 and 22% of those who say they have limited experience and usually follow instructions expect AI to have only a negative impact on their job in the next five years, compared with about 15% of workers ages 25 to 64 and 9% of those who set strategy for their organization.
-
Workers who use AI are far more positive than other workers about AI’s effect on their line of work. Among workers who use AI at least once a month, 51% say it is currently making people who do work like theirs more productive, compared with 10% of other workers.
-
The more people who use AI at a workplace, the more workers say their employer expects of them. Among workers who say almost everyone at their workplace uses AI (11% of workers asked), 66% say their employer has expected more work from them since AI became available. Where workers say only a few people use AI (33% of workers asked), 25% say so.
-
AI users with formal employer training are more likely to save time. Among workers who use AI, about a third say their employer provided formal training. Of these, 83% say AI saves them at least an hour a week, compared with 67% of users without formal training.
-
Time savings differ widely across industries. Three in four workers in information / technology (74%) say AI saves them at least an hour a week, compared with 26% of those in health care and social assistance and 20% in hospitality or service. Banking and finance (54%) and professional services (50%) come next; the remaining industries fall between 22% and 37%.
This report is organized in four parts: (1) how often workers use AI and how much time they say it saves; (2) how the saved time is distributed, whether employers now expect more, and who workers think benefits most; (3) which workers save time, by industry, gender, education and income, age, job level and tasks, race and ethnicity, and employer training; and (4) whether workers think AI is making, or will make, people in their line of work more or less productive.
This report is part of a series from Northeastern University’s AI & Society Survey. Earlier reports include Divided on AI, Wary of Its Power (https://www.nuinfoinstitute.org/publications-1/divided-on-ai-wary-of-its-power) and AI in the Hiring Room (https://www.nuinfoinstitute.org/publications-1/ai-in-the-hiring-room).
-
22% of Americans would support a new AI data center in their own community; 54% would not
-
70% of Republican men with graduate degrees would support a local data center — the standout subgroup in the education-by-gender comparison
-
Within 4 percentage points between Democrats and Republicans on four AI-policy questions: job retraining, lifelong learning accounts, exports of advanced AI, and the influence of tech leaders
-
66% of independents do not trust AI companies — the highest level of distrust among Republicans, Democrats and independents
About the data. The Civic Health and Institutions Project (CHIP50) conducts repeated large-sample national surveys of
Americans on health, politics, and civic life. This report draws on the Northeastern University AI & Society Survey, fielded
through CHIP50 from June 21 to August 11, 2026, with 38,375 responses from adults 18 and older in all 50 states and the
District of Columbia. The surveys were fielded through PureSpectrum via an online, nonprobability sample, with state-level representative quotas for race/ethnicity, age, and gender (for more details, see chip50.org and covidstates.org). In addition to balancing on these dimensions, we reweighted our data using demographic characteristics to match the U.S. population with respect to race/ethnicity, age, gender, education, and living in urban, suburban, or rural areas.
All percentages are weighted. Which respondents were asked each question, and how the report compares groups, is
described under Detailed methods in the appendices.
How this report was made. AI was used throughout the production of this report to query CHIP50 data through the Tech & Society connector and the survey’s published weighted tables, compute figures, draft text, and build charts. The AI models used were from Anthropic’s Claude Opus series. Each figure is traced to its source. The authors chose the questions, interpreted the results, and every figure was independently checked against the source tabulation by at least one author. The authors reviewed the report and are accountable for what it contains.
About IIID. The Institute for Information, the Internet & Democracy (IIID) at Northeastern University advances publicinterest research that examines how the internet and emerging information technologies impact democratic societies.
Funding. Northeastern University provided funding for this survey.
CHIP50 is a collaboration of Northeastern University, Harvard University, Rutgers University, University of Rochester, and Massachusetts General Hospital.
Contact. infoinstitute@northeastern.edu
methods
Nearly half of American workers (45%) use AI for work at least once a month. About one in eight (13%) use it several times a day. At the other end, 10% have tried AI only once or twice and 45% have never used it for work.
A small slice of workers accounts for most AI use at work. The survey asked how often workers use AI, not how many times, so this can only be estimated. Under a range of conservative assumptions about how many uses a week each answer stands for, the 13% of workers who use AI several times a day account for roughly 70% to 78% of all AI use at work. The 20% who use it at least once a day account for 85% to 88%.

Method note: The estimate of how concentrated AI use is multiplies the share of workers in each frequency group by an assumed number of uses per week and then compares each group’s total with the total for all workers. The assumed uses per week are: none for those who have tried AI only once or twice; about one a month for monthly users; one for weekly users; two to four for those who use it multiple times a week; five to 7.5 for those who use it once or twice a day; and 12.5 to 30 for those who use it several times a day. In three scenarios that pair the low, middle and high values for the three heaviest groups (two, five and 12.5 uses a week; three, 7.5 and 20; four, 7.5 and 30), the share of all use from workers who use AI several times a day ranges from 70% to 78%, and the share from all daily users from 85% to 88%. The estimate counts how often workers use AI, not how much work each use involves.
Most workers who use AI say it helps their work, and fewer say it hurts. Among AI users, about half (51%) say AI makes their job easier or less stressful, and 43% each say it improves the quality of their work or increases their productivity (Figure 2). Fewer report harm: 13% say AI makes their job harder or more stressful, 12% say it reduces the quality of their work and 8% say it decreases their productivity. About a quarter (24%) see no significant benefit or harm.
This report uses self-reported hours saved as its main measure of AI’s benefits. It is an incomplete measure. It leaves out other benefits workers report, such as better work or less stress; it does not factor in any time AI might cost workers, which the survey did not ask about; and it rests on workers’ own estimates.
Among workers who use AI, 72% say it saves them at least an hour a week, or a third of all workers. The survey first asked AI users how AI has affected their work, letting them select all answers that applied. Those who chose at least one benefit (AI makes the job easier or less stressful, improves the quality of their work, or increases their productivity) were then asked how many hours a week AI saves them compared with their work before they used AI tools. About three in four users (76%) were asked; nearly all of them name at least one hour, and 2% of users say zero hours. The rest of the users reported no benefit and were not asked.
Among workers who use AI and say it benefits their work, 63% say it saves them at least four hours a week, or half an eight-hour workday, and 26% say eight hours or more, a full workday. The median answer is five hours. The top answer, 20 hours or more, is given by 3%.

Method note: Respondents could select more than one answer, so the shares add to more than 100%. Only those who selected one of the first three benefits were asked how much time AI saves them, so this question is the gate for every time-saved figure in this report.
Workers who use AI more often are generally more likely to save time, and they save more of it. Among those who use AI several times a day, 83% say it saves them at least an hour a week, compared with 77% of those who use it multiple times a week, 75% of those who use it once or twice a day, 60% of weekly users and 50% of monthly users. Averaged over everyone in each group, those who use AI several times a day save 6.3 hours a week; monthly users save 2.0 hours.

who uses AI at work and how much time it saves
The time AI saves goes mostly to a small share of workers. Half of all the hours saved go to fewer than one in 10 workers (8%), and the 10% who save the most time get 58% of them. The other 67% of workers either do not use AI regularly or report no time saved. Weighting each hour by household income, a rough measure of what that time is worth, sharpens the pattern: half of the value goes to about 6% of workers, and the 10% whose saved time is worth the most account for 68% of it.


At workplaces that use AI, workers who save time with AI are much more likely than other AI users to say their employer now expects more work from them. Among AI users at these workplaces, 57% of those who say AI saves them at least an hour a week say their employer has expected more work from them since AI became available. Among other AI users, 32% say so, and a majority (63%) say expectations have stayed the same.
The more widely AI is used at a workplace, the more workers say their employer expects more of them. The survey asked most workers (17,851 of the 20,275) whether AI is used in their workplace. About a third (36%) say it is not used at all, 33% say a few people use it, 20% say many people do and 11% say almost everyone does. Workers at workplaces that use AI, whether or not they use it themselves, were then asked about their employer’s expectations. Where only a few people use AI, 25% say their employer has expected more work from them since AI became available. That share rises to 53% where many people use it and 66% where almost everyone does. Fewer than 5% in each group say expectations have decreased. Workers at workplaces that do not use AI were not asked, so they cannot be compared.
These are workers’ own reports. Workers who save time may differ from other users in their jobs and employers.
Method note: The expectations question was asked of workers asked about their workplace who said AI is used there. The comparison
covers the 7,310 AI users who answered it: 5,361 who save at least an hour a week and 1,949 other users, including those who reported no benefit and were not asked about time saved.
More workers see AI’s productivity gains as shared than say either employers or workers get most of them. Among workers at these workplaces who say AI helps people there get more done, 32% say mostly employers (through higher profits or reduced headcount), 24% say mostly workers (through higher pay, fewer hours, or better conditions), and 40% say both benefit about equally; 4% are not sure. Views track how much workers think AI helps. Among those who say it helps the people at their workplace get a lot more done, 23% say employers benefit most and 30% say workers do. Among those who say it helps them get a little more done, 39% say employers and 18% say workers.
Views also differ with workers’ own use of AI. Among workers at these workplaces who never use AI, about half (49%) say employers benefit most, and 16% say workers do. Among those who use AI several times a day, 23% say employers and 30% say workers. AI users who save at least an hour a week split nearly evenly (28% say employers, 26% workers), and 44% say the benefits are shared about equally. Among other monthly AI users, 36% say employers benefit most and 21% say workers.

where the saved time goes
Workers in information / technology are about three times as likely as those in health care
and social assistance to save time with AI. Among all workers, 74% of those in information / technology say AI saves them at least an hour a week, compared with 26% in health care and social assistance and 20% in hospitality or service. Banking and finance (54%) and professional services (50%) come next; the other industries fall between 22% and 37%. On average,
information / technology workers save 5.3 hours a week, counting everyone in the industry; health care workers save 1.3 hours.
Part of this gap comes from who uses AI, but it is also reflected among users of AI in each sector. Among AI users, 86% in information / technology save time, compared with 69% in other industries, and 64% in health care, compared with 73% in other industries. In each of the five frequency groups, a smaller share of health care users than of users in other industries save time; among those who use AI several times a day, 75% of health care users do, compared with 84% in other industries.

Among all workers, men are more likely than women to save time with AI. Among working men, 38% say AI saves them at least an hour a week, compared with 27% of working women. Averaged over all workers, men save 2.3 hours a week and women 1.4. Part of the gap reflects use: half of men (50%) use AI for work at least once a month, compared with 40% of women. But a gap persists among AI users, too: 75% of male users and 67% of female users save time. Among workers who report a benefit, men also put the time saved higher: an average of 6.0 hours a week, compared with 5.1 hours for women, counting the answer “20 or more” as 20 hours.
Time saved with AI rises steeply with both education and household income. Among workers with a graduate degree, 59% say AI saves them at least an hour a week, as do 40% of college graduates, compared with 25% of those with some college, 20% of high school graduates and 14% of those who did not finish high school. Averaged over all workers, those with a graduate degree save 3.6 hours a week, compared with 1.2 hours for high school graduates. By household income, the share is 19% in households earning under $50,000, rises to 42% at $75,000 to $100,000 and reaches 54% at $100,000 or more. Part of the gap reflects use: 74% of workers with a graduate degree and about half of college graduates (54%) use AI for work at least once a month, compared with 22% of those who did not finish high school; by income, 30% of workers
in households earning under $50,000 use it, compared with 69% at $100,000 or more. Among AI users, education still matters: 80% of users with a graduate degree and 74% of college graduates save time, compared with 67% of those with some college or a high school diploma and 63% of those who did not finish high school. Among users, 65% of those in households earning under $50,000 save time, compared with 79% at $100,000 or more.

Workers ages 25 to 44 are the most likely to save time with AI. Among workers ages 25 to 44, 39% say AI saves them at least an hour a week, compared with 28% of those 45 to 64, 25% of those 18 to 24 and 19% of those 65 and older. Use follows the same pattern: 52% of workers ages 25 to 44 use AI for work at least once a month, compared with 37% of those 18 to 24, 41% of those 45 to 64, and 29% of those 65 and older. Among AI users, 75% of those 25 to 44 save time, compared with 67% to 69% in the other age groups.
The youngest workers are also the most likely to see AI as a threat: 27% of workers ages 18 to 24 expect AI tools to have only a negative impact on their own job in the next five years, compared with 15% to 16% of older workers.

Among all workers, those higher in their organizations are far more likely to save time with AI. Some 14% of workers who say they have limited experience and usually follow instructions say AI saves them at least an hour a week, compared with 53% of those who manage a department and 60% of those who set strategy for their organization. Averaged over all workers at each level, those who set strategy save 4.1 hours a week; those with limited experience, 0.8 hours. Part of the gap reflects use: 22% of workers with limited experience use AI for work at least once a month, compared with 68% of department managers and 74% of those who set strategy. Looking only at AI users does not close the gap: 63% of users with limited experience save time, compared with 67% of those who manage their own projects, 73% of team managers, 78% of department managers and 81% of those who set strategy. Owners (76%) and freelancers or contractors (73%) also save time more often than the least experienced users.
The least senior workers also see more risk. Among all workers, 22% of those with limited experience expect AI tools to have only a negative impact on their own job in the next five years, compared with 10% of department managers and 9% of those who set strategy.
What a job involves matters, too. Among all workers, 66% of those whose jobs involve making budget or software purchasing decisions say AI saves them at least an hour a week, as do 57% of those whose jobs involve data processing or analysis, 53% of those whose jobs involve marketing or content creation, and 46% of those who manage or supervise others. Among workers whose jobs involve none of these four tasks, 17% do. Among AI users, 84% of those who make purchasing decisions save time, compared with 69% of those who do not. Working directly with clients or customers makes little difference among users (72% vs. 73%).


Unlike gender, education, and income, race and ethnicity have little association with whether workers save time with AI. Among all workers, 33% to 38% in each racial and ethnic group say AI saves them at least an hour a week: 38% of Asian American workers, 35% of African American workers, 33% of Hispanic workers and 33% of White workers. Use is similar, too: 45% to 51% of workers in each group use AI for work at least once a month, with Asian American workers the most likely to do so. Among AI users, 72% to 73% in each group save time.

Among workers who use AI, those whose employer provided formal training are more likely to save time with AI. Of users with formal training, 83% say AI saves them at least an hour a week, compared with 67% of users without it. Formal training means workshops, courses, or seminars; users without it include those who received only informal guidance, those who received none, and those who are not sure.
Trained workers also use AI more often, but the gap appears within each of the five frequency groups. Among workers who use AI several times a day, 86% of those with formal training save time, compared with 78% of those without; among monthly users, 67% compared with 46%. Employers may offer training where AI is most useful, so the gap does not show that training causes the time savings.
Formal training is unevenly spread, largely in the same pattern as who saves time. Among all workers, 14% use AI for work at least once a month and say their employer provided formal training in it. That includes 17% of working men and 10% of working women; 32% of workers with a graduate degree, compared with 4% to 16% of those with less education; 29% of those in households earning $100,000 or more, compared with 6% of those earning under $50,000; and 18% of urban workers, compared with 9% of rural workers. Workers ages 25 to 44 are the most likely to have had formal training (18%) and those 65 and older the least likely (4%). The same groups lead among AI users: 38% of men and 29% of women who use AI report formal training, as do 47% of users with a graduate degree and 46% of users in households earning $100,000 or
more. Most of these differences hold when other traits are taken into account (for education, only graduate-degree holders stand out); with those traits considered, African American users are also somewhat more likely, and Asian American users somewhat less likely, to report formal training. The groups that get formal training most are largely the ones that save the most time, one reason the link between training and time saved may partly reflect differences in jobs and employers.

Method note: The training question was asked of workers asked about their workplace. “Formal training” is the answer “Yes, formal
training (e.g., workshops, courses, seminars).” The comparison of time saved covers the 8,031 AI users who answered the training question. In Figure 13, the 1,034 AI users who were not asked the training question or skipped it are counted as not having formal training, so those shares may slightly understate training.
who saves time — differences by industry and demographics
Asked about people who do work like theirs, more workers say AI is making them more productive than less productive, and about half see no change. Today, 29% of workers say AI is making people who do work like theirs more productive, 20% say less productive, and 51% say it is making no change. Looking ahead, 31% expect AI eventually to make people in their line of work more productive, 23% expect less, and 46% expect no change.
These questions ask about other people in the same line of work, not about the respondent, and they were asked of all workers, including those who never use AI.
Workers who use AI are far more positive about their line of work than those who do not. Among workers who use AI at least once a month, 51% say it is currently making people who do work like theirs more productive, and 14% say less productive. Among other workers, including those who have only tried AI, 10% say more productive and 25% less; most (65%) see no change. The gap holds looking ahead: 52% of users and 14% of other workers expect AI eventually to make people in their line of work more productive. These views may reflect the jobs people hold as much as their own experience, since workers in jobs where AI helps are more likely to use it.

what workers expect
Download the full report with appendices.
citation
Lazer, D., Uslu, A. A., Wihbey, J., Zomer, A., Bernt, S., Howard, J. X., Li, C. Z., Waldrip, C., & Bader-Gregory, E. (2026). AI and Time at Work: For a Few Workers, Most of the Hours Saved (IIID Research Report). Institute for Information, the Internet & Democracy, Northeastern University. https://www.nuinfoinstitute.org/publications-1/ai-and-time-at-work.
topics
Technology & Work
Artificial Intelligence
related work

Measuring Agency-Amplifying AI Literacy: The HuskyAI Coaching Platform for Workforce Readiness
July 2026
HuskyAI is a web platform on which students develop AI knowledge by working through authentic
Artificial Intelligence
Artificial Intelligence
Working Paper


Measuring Agency-Amplifying AI Literacy: The HuskyAI Coaching Platform for Workforce Readiness
July 2026
HuskyAI is a web platform on which students develop AI knowledge by working through authentic
Artificial Intelligence
Artificial Intelligence
Working Paper


Measuring Agency-Amplifying AI Literacy: The HuskyAI Coaching Platform for Workforce Readiness
July 2026
HuskyAI is a web platform on which students develop AI knowledge by working through authentic
Artificial Intelligence
Artificial Intelligence
Working Paper

.png)