A representative survey by Epoch AI and Ipsos found that 20 percent of employed Americans now delegate at least one work task to AI that was previously handled by coworkers or outside contractors.
The study interviewed 1,106 employed US adults between July 10 and 19, 2026. Respondents were asked about ten common work tasks drawn from US Department of Labor data. Workers reported using AI across all ten tasks, though adoption rates varied widely.
Software development leads, record-keeping trails behind
AI use is highest in computer systems and software development, where 57 percent of workers who do that task use AI. Data analysis comes in at 46 percent, and reading work documents at 39 percent. Record-keeping sits at the bottom, with just 25 percent of affected workers using AI.
But using AI does not mean the machine takes over the whole task. Workers mostly describe AI as partial support. Full or near-full task completion by AI hits 10 percent only in software development and stays below 7 percent for everything else.
The shift from human to machine work shows up most clearly in data analysis: 7.1 percent of respondents say AI has taken over tasks that people used to handle. Reading work documents follows at 5.7 percent, then record-keeping at 5.3 percent. Epoch AI stresses that this task substitution does not automatically mean workers are being displaced entirely.
More AI use often saves time, but not always
The survey shows a link between how much AI does and how much time workers say they save. When AI handles only part of a task, respondents report saving time on 37 percent of those tasks. When AI does most or all of the work, that figure rises to 53 percent.
Whether heavier AI use actually makes people faster or whether workers simply lean on AI more when they want to save time cannot be determined from the data alone.
AI does not always speed things up, either. About one in six AI-assisted tasks now takes longer than before, according to the survey. The researchers suggest this may be because interacting with AI itself eats up time or because workers use the freed-up capacity to do tasks more thoroughly or at higher quality.
Two-thirds of AI output gets used with little editing
Another finding concerns how workers handle what AI produces: 66 percent of AI output gets used unchanged or with only minor tweaks. Just 6 percent is used without any changes at all. On the other end, 5 percent of output gets heavily reworked or mostly rewritten.
The researchers note that low editing effort is not a direct measure of AI output quality. They also found no consistent link between reported time savings and how much editing workers did.
Tasks are shifting between humans and AI
Epoch AI sums up the findings by calling AI a “versatile but usually not self-sufficient workplace tool.” The data points more toward a redistribution of tasks between humans and AI than toward full automation of entire jobs.
The results are based on self-reported data. Neither actual time savings nor the quality of AI output were measured objectively. The tasks studied were selected based on national employment data, not on how likely they are to be affected by AI.
Earlier surveys point in the same direction
A Gallup survey from August 2025 found that 45 percent of US workers use AI on the job, though only 10 percent use it daily. A recent Anthropic survey of roughly 9,700 Claude users showed that about half of respondents believed AI could already handle 50 percent or more of their work. That sample consisted of users of a specific AI product, though, not a representative cross-section of the population.
What it means
For people doing the work, the change is practical rather than theoretical. Roughly one in five workers is already moving specific duties away from colleagues and toward software. This happens mostly in coding and analysis. Most people still treat AI as a helper rather than a replacement. The time savings are real in some cases, but the extra time spent talking to the tool cancels out gains in others. Workers are also trusting the output more, editing it less than before. The shift is about moving tasks around, not wiping out jobs.
