Contrary to expectations that AI would transform office workers' jobs, research findings show fatigue and burnout are rising in reality, TechCrunch reported on Feb. 10.

A study published in Harvard Business Review found that adopting AI accelerates burnout rather than reducing workloads. Researchers at UC Berkeley recently observed an IT company of about 200 employees that actively adopted AI over 8 months and conducted more than 40 in-depth interviews.

The findings showed that as productivity rose thanks to AI tools, workloads also increased naturally and work expanded into lunch and evening hours. Rather than creating spare time as expected, higher productivity from AI increased workloads in reality.

One engineer said, "AI lets me do more work, but it ultimately made me work longer." A reaction on Hacker News also said, "After adopting AI, expectations for the team tripled, but actual productivity rose only 10 percent."

The researchers warned that "AI is not supporting employees' work but is instead increasing fatigue and stress." Expectations that AI would bring workplace innovation are producing the opposite result in reality.

Separately, results from last summer from the National Bureau of Economic Research (NBER), which tracked AI adoption across thousands of workplaces in the United States, showed productivity gains were only 3 percent in terms of time saved and did not have a meaningful impact on income or working hours in any occupation.

The researchers said AI strengthens employees' work capabilities, but pointed out that the result is leading to fatigue and burnout and to higher organisational speed and expectations for responsiveness.