The work tasks people most want AI to take over

Workers are more open to automation than the headlines suggest, but they are picky about which tasks. In the Stanford WORKBank study, workers expressed a positive attitude toward automating about 46% of all studied tasks. The top reason, cited about 69% of the time, was to free up time for high-value work. People do not want to be replaced. They want the busywork gone.

What the study measured

The Stanford WORKBank study surveyed 1,500 US workers and 52 AI experts across 844 work tasks in 104 occupations, collected in early 2025. It builds on the US Department of Labor O*NET task list.

Each task carries two scores. Workers rate their desire to automate it from 1 to 5. Experts rate AI capability from 1 to 5. We call a task wanted when workers rate desire 3 or higher, and AI-capable when experts rate capability 3.5 or higher. The gap between these two scores is where the interesting story lives.

Almost half of tasks are welcome to go

Across all 844 tasks, workers showed a positive attitude toward automating about 46% of them. That is close to half. It runs against the common assumption that workers resist automation across the board. When a task drains time without using their skills, most people are happy to let it go.

The reasons are practical. The most common, named about 69% of the time, was to free up time for high-value work. Another frequent reason was that the task is simply repetitive. Workers can tell the difference between work that uses their judgment and work that just fills the day.

What people want automated

The pattern is consistent. Workers welcome automation of tasks that are:

• Repetitive and predictable
• Low in judgment, with a clear right answer
• Time draining without being rewarding
• Standing between them and the work they care about

These are the tasks that, once gone, give people room to do the parts of the job that matter.

What people want to keep

The resistance is just as clear. Workers tend to hold on to tasks that involve:

• Judgment and weighing trade-offs
• Relationships and trust with people
• Accountability, where someone must own the outcome

These tasks define the human part of a role. People do not want to hand them over, even when they could. That instinct lines up with where AI is weaker too, which is reassuring.

Wanted is not the same as AI-capable

Here is the catch. A task can be wanted but not yet doable, or doable but not wanted. Desire and capability are two separate scores, and they do not always agree. To make sense of the four combinations, the study sorts every task into a zone.

Zone	Workers want it	AI can do it	What it means
Green Light	Yes	Yes	Automate now, with confidence
R&D Opportunity	Yes	Not yet	Wanted, but the tools need to catch up
Red Light	No	Yes	AI could, but people would rather it did not
Low Priority	No	No	Leave it for now

The Green Light zone is the sweet spot. These are tasks workers want gone and AI can already handle. There is no conflict and no waiting. Of the 844 tasks studied, 281, about 33%, fall into Green Light. That is roughly one task in three where automation helps people and the technology is ready today.

R&D Opportunity tasks are worth watching. People want them automated, but the capability is not there yet. As tools improve, these are the tasks likely to move into Green Light next. Red Light tasks are a caution. Just because AI can do something does not mean it should, especially when workers value keeping it.

A note on honesty

These figures are aggregates across many workers and tasks, not a prediction about your specific role. The capability ratings are expert opinion, not a benchmark of any single product. And the data is a 2025 snapshot, so the picture will move as both attitudes and tools change.

What to do with this

If you manage a team or shape your own work, start with the Green Light tasks. They carry the least friction: people want them automated and the tools can deliver. Clearing them frees time for the judgment, relationship, and accountability work that people want to keep and that AI cannot cover.

You can see the full set of Green Light tasks and which roles they touch. Explore the Green Light tasks to find the safe, wanted automation wins first.

The bottom line

Workers welcome automating about 46% of tasks, mostly to free up time for high-value work. They want repetitive, low-judgment tasks gone and want to keep work that needs judgment, relationships, and accountability. The clearest place to start is the Green Light zone, the 281 tasks people want automated that AI can already do.