Fed survey: one in four workers used generative AI at work
One in four US workers used generative AI at work in the prior month, and 81% of those users agreed it saves them time, according to the Federal Reserve Board's Economic Well-Being of U.S. Households in 2025 report, released on May 13, 2026. The survey behind it was fielded in October 2025. For employers, the finding shows both the reported payoff and its limit: workers who use the tools see benefits, while those who have not tried them see fewer.
What the Federal Reserve report measured
The report is the Federal Reserve Board's yearly look at the finances of US adults and their families. Economic Well-Being of U.S. Households in 2025 draws on the Board's annual Survey of Household Economics and Decisionmaking, known as SHED, which was fielded in October 2025. It covers financial well-being, employment, income and expenses, and housing.
This edition added a measure of something employers now ask about: whether workers use generative artificial intelligence in their jobs. The Board released the report on May 13, 2026, together with a fact sheet, downloadable data and data visualizations.
How many US workers use generative AI on the job
The headline answer is a minority. One in four workers used generative AI at work in the month before the survey. That is a measure of recent use, not of access: it counts workers who actually used the tools in the prior month.
Among those users, the reported benefit is time. Eighty-one percent of users agreed that generative AI saves them time. The figure comes from workers' own answers, so it describes what users believe about their work rather than a measured change in output.
How users and non-users see AI differently
The survey found a split in attitudes that follows use. AI users were more likely to agree that the technology would improve their careers than to say they were worried that AI would replace their jobs. Workers who had not used AI in the prior month saw fewer potential benefits.
The report does not say which way the effect runs. Workers who expect benefits may be the ones who try the tools, or trying the tools may change what workers expect. Both readings fit the published numbers.
What the job market looked like when the survey ran
The AI findings arrived in a labor market the Board described as solid but softer than a year earlier. Forty-two percent of adults said finding or keeping a job was a minor or major concern, up from 37 percent in 2024.
The share of adults who voluntarily left a job declined slightly to 8 percent. Layoffs rose a little: 7 percent of all adults reported being laid off, up from 6 percent in 2024.
How households were doing overall
Household finances were broadly steady. The report shows 73 percent of adults were doing okay or living comfortably financially, the same as in 2024 but below the high of 78 percent in 2021. The share who would cover a $400 emergency expense with cash or its equivalent was unchanged at 63 percent.
Prices remained the most common worry. The share citing them as a major or minor concern held at 91 percent, while the share calling price increases a major concern fell to 53 percent from 56 percent in 2024.
"As we work to support a strong and vibrant economy, it's critical for the Federal Reserve to understand the economic experiences of families and communities." — Michael S. Barr, Federal Reserve Board Governor
What it changes for employers rolling out AI tools
For employers, the survey answers two practical questions about adoption. First, use is still limited: three in four workers had not used generative AI at work in the prior month. Second, the people who do use it mostly report saving time, which is the task these tools are bought for.
The gap between users and non-users matters for any rollout. Workers who had not tried the tools saw fewer benefits, so an employer counting on the time savings users report would be relying on a group that is, for now, the smaller one. The survey gives no figures on how much time is saved or on output, so it cannot say what the tools are worth in dollars.
Questions and answers
What share of US workers use generative AI at work?
According to the Federal Reserve's survey, one in four workers used the technology on the job in the prior month. The survey was fielded in October 2025.
Does generative AI save workers time?
Most users say it does. Eighty-one percent of workers who used generative AI agreed that it saves them time. The report does not measure how much time.
Are workers worried that AI will replace their jobs?
AI users were more likely to say the technology would improve their careers than to say they worried it would replace their jobs. Non-users saw fewer potential benefits.
MOZGII view
How we would build it
A survey like this tells a company where it stands before it spends on tools. Most workers have not used generative AI at work in the past month, and the ones who have mostly say it saves time. That is a useful starting point, but a belief about saved time is not the same as saved time, and we would treat it that way.
If we were rolling out an AI assistant inside a company, we would pick two or three tasks where time is easy to count, such as drafting replies, summarizing documents or filling forms. We would record how long each task takes before the tool and after it, for the same people. Team leads would check a sample of the output each week, because a fast draft that needs heavy rework saves nothing. Access would follow the data the task needs, and nothing more.
It would break in the gap the survey describes. People who already use the tools are keen; people who do not see little point, and they are the larger group. A rollout that only counts enthusiasts would look better than it is. The one metric we would watch is minutes saved per task across the whole team, users and non-users together.
Next, we would build short, task-specific guides for the people who have not tried the tools, since they are where most of the untapped time would sit.
Sources
Where the facts in this story come from.