New in AI

◆ AI AT WORK · Adoption

The government asked workers what AI saves them. The most common answer was an hour or two a week.

The News

About 55% of US workers said they have used AI on the job for at least one of 11 tasks, according to the Census Bureau's March 2026 Household Trends and Outlook Pulse Survey, published on August 11, 2026. Workers who had used AI at work in the last week were asked how many additional hours they would have needed without it: 31% said one to two hours, 25% said less than an hour, 15% said three to four hours and 15% said more than four. Another 10% said it saved no time and 3% said using it required additional time. The most common uses were searching for information or technical help (37%), writing communications or documentation (32%) and generating ideas (32%). All comparisons were tested at the 90% confidence level.

Chandni Melwani

Chandni Melwani

Founder & Editor

Aug 14, 2026 · 2 MIN READ

A workshop wall rack holding rows of chisels, hand saws, screwdrivers and wooden planes above a workbench US Photo: Barn Images / Unsplash

Know More

  • — The frequency: among workers who use AI on the job, 24% used it every day in the last week, 46% used it at least one day, and 30% did not use it at all that week.
  • — The education gradient: 46% of workers with a bachelor's degree or higher used AI to interpret, translate or summarise information, against 25% with some college and 16% with a high school diploma or less.
  • — The usage gap: 75% of degree-holders used AI at work on at least one day last week, against 59% of those with high school or less.
  • — The sex gap: 30% of male workers used AI every day last week, against 17% of female workers.
  • — The method: fielded in March 2026 by the Census Bureau's Household Trends and Outlook Pulse Survey, which the Bureau says transitioned into a cross-sectional design that same month, after running longitudinally through 2025. Time savings are self-reported.

Ten per cent of workers who used AI at work last week said it saved them no time at all. Another 3% said it cost them time. Those two lines, from the Census Bureau’s March 2026 pulse survey, are the ones a vendor deck never carries.

The headline number is friendlier. About 55% of US workers have used AI on the job for at least one of 11 tasks, and of those who used it in the past week, 31% said it saved them one to two hours and 25% said under an hour. Set against a 40-hour week, the typical return is low single-digit percentages of time. Fifteen per cent reported three to four hours and another 15% more than four, so the upside is real, but it sits at the tail.

The survey is worth reading for what it measures rather than what it concludes. Workers were asked how many additional hours the same work would have taken without AI, which is a narrower and more answerable question than whether AI transformed their role. It is self-reported, the answers were collected in March and published in August, and no federal agency is selling a seat licence.

The gradient underneath is the part with consequences. Among degree-holders, 46% used AI to interpret, translate or summarise information, against 16% of workers with a high school diploma or less, and 75% used it on at least one day last week against 59%. Adoption is currently concentrated where the analytical work already was.

Which is a different picture from the deployments that make headlines. ADNOC running AI across 120 drilling rigs and Shopify reporting AI orders up 13x describe systems rebuilt around AI. This describes 55% of workers finding an hour.

Related

#AI Adoption#Future of Work#Productivity#Census Bureau#Enterprise AI#United States

Frequently Asked Questions

Does 55% mean most American workers use AI regularly?

No. It means 55% said they have used AI on the job for at least one of 11 listed tasks at some point. Among those users, 30% had not used it at all in the previous week and 24% had used it every day. Ever-use and habitual use are different measures, and the gap between them is most of the story.

Why does an hour or two a week matter if it sounds small?

Because it is a defensible baseline. Vendor surveys usually report perceived transformation; this asks how many extra hours the work would have taken and records the answers, including the 10% who saved nothing and the 3% who lost time. At roughly one to two hours against a 40-hour week, the honest claim is a few per cent of time recovered, not a reinvented job.

What is HTOPS?

The Household Trends and Outlook Pulse Survey, a Census Bureau survey of social and economic conditions in US households, which moved to a cross-sectional design in March 2026 after running longitudinally through 2025. Its comparisons are significance-tested at the 90% confidence level. The AI figures appear only in the Bureau's August 11 article: the published March 2026 table package covers health, spending, housing, transport and the arts, and contains no AI tables.

Sources

Chandni Melwani

Chandni Melwani

Chandni Melwani is the founder and editor of New in AI, covering AI agents, M&A, and enterprise adoption. She holds a Master's in Management of Artificial Intelligence from Queen's University and brings a practitioner's perspective from her work in Data and AI leadership.

Follow ↗

AI at Work Briefing

Get the five-beat briefing in your inbox.

The big story, frontier labs, AI at work, e-commerce and the capital crossing borders, each with the part that tells you why it matters, not just that it happened.

Free · No spam · Unsubscribe in one click

More from AI at Work