Artificial intelligence and the productivity paradox: why efficiency can mean more work

The idea seems simple: if a task that used to take an hour can now be done in 20 minutes, you’ve got 40 extra minutes in your day. But the time saved doesn’t always translate into rest outside of work.

The opposite can happen, too. When a task becomes faster, it creates space for another task, and then another. The expectation of productivity increases, and what used to take up a morning becomes just the start of a growing to-do list.

In this context, researchers have observed a curious relationship between artificial intelligence, working hours, and well-being.

When productivity turns into more work

A study conducted by researchers from Emory, Auburn, and Fordham universities analyzed nearly two decades of data from the American Time Use Survey, linking AI exposure in various professions to the number of hours worked.

Between 2004 and 2023, occupations with higher exposure to technology saw longer working hours. The difference between workers in the 25th and 75th percentile of AI exposure reached up to 2.2 additional hours per week.

The period following the launch of ChatGPT produced an even more striking result. Professionals in roles with greater exposure to generative AI began working about 3.15 hours more per week compared to those in less exposed positions.

The research does not claim that using ChatGPT alone causes longer working hours. The findings point to an association and help explain why productivity gains may not necessarily lead to more free time.

The time saved is quickly filled

When a tool speeds up a task, companies may expect more work to be completed in the same amount of time. For those working with performance goals, there’s another effect: producing more can become a new standard for evaluating success.

The change is subtle. Previously, writing five reports a week might have been considered a good performance. If AI allows you to produce those five reports in half the time, expectations may shift to ten.

The pressure can come from monitoring tools themselves

Another mechanism observed by researchers involves digital monitoring tools. These tools can track productivity, activity time, and performance indicators, especially in remote or hybrid work environments.

When this monitoring translates into automated scores or targets, technology can increase the pressure to maintain pace or avoid falling behind colleagues.

A systematic review published in October 2026 reinforces that the effects of AI on burnout vary depending on how it is used. Tools that reduce tasks and cognitive load can decrease exhaustion, while monitoring and algorithmic control are linked to increased pressure and fatigue.

Therefore, referring to this phenomenon as “AI burnout” requires caution. The term describes a possible experience of work mediated by technology but does not imply that AI has created a new medical category of burnout.

What happens when the workday doesn’t decrease

The central question becomes who benefits from the time saved. If a task becomes faster, this gain can mean less work, more production, cost reduction, or a combination of these possibilities.

When the choice is always to increase the volume of deliverables, technology can improve productivity without improving quality of life.

For those who use AI at work, this creates an important distinction between doing things faster and working less.