AI tools were sold as a way to free up time. Two 2026 studies suggest that, without deliberate management, they often do the opposite.
Work gets more intense
Aruna Ranganathan and Xingqi Maggie Ye spent eight months studying a US tech company of about 200 people. They published the findings in Harvard Business Review in February 2026.
Employees worked faster, took on a broader range of tasks and stretched work into more hours of the day. Often nobody asked them to.
AI let people step into work that used to belong to others. They ran several threads at once, with agents working in parallel. The researchers warn this leads to fatigue, burnout and weaker decisions.
"Brain fry"
In March 2026, BCG researchers described "AI brain fry" in Harvard Business Review. They surveyed 1,488 US workers.
Heavy oversight of AI was linked to 14% more mental effort and 12% more mental fatigue. It also brought 19% more information overload.
People with brain fry were more likely to plan to quit: 34% compared with 25%. When managers gave training and support on AI tools, brain fry fell.
Build this capability in your organisation
Growth Performance designs bespoke programmes that turn insight into measurable behaviour change.
Managers are stretched too
Gallup's 2026 report found manager engagement fell from 27% to 22% in a year. The people expected to lead AI change are among the most depleted.
What managers can do
The Berkeley researchers recommend building an "AI practice": shared norms for how and when AI is used.
Batch AI work. Keep AI-heavy tasks to set blocks of the day. Protect time for focused work without it.
Pause on purpose. Add short, planned breaks before big decisions. Overload builds quietly.
Keep human contact. Protect time for listening and conversation. Constant solo work with AI can isolate people.
Remove before you add. Every new AI task needs something taken away. That is part of Adaptation: people directing and checking AI in real work, with time made for it.
Try this
Ask your team how many hours a week they spend checking AI output. Compare that with the time AI saves them.
The Checking Capacity Calculator shows whether your team has room to check properly. The Change Load Counter shows how much change they are carrying.