We’ve all seen the claims: AI makes our work easier and is going to boost productivity to unheard of levels. For example, according to PwC, AI could boost global output by 15% over the next decade.

But maybe things are not quite as rosy as the AI cheerleaders would have us believe. Some new research from Harvard Business Review (HBR)has suggested that AI benefits have been somewhat overhyped.

The research found that employees using AI tools worked faster and for longer hours, juggling more tasks. “These changes can be unsustainable, leading to workload creep, cognitive fatigue, burnout, and weakened decision-making. The productivity surge enjoyed at the beginning can give way to lower quality work, turnover, and other problems,” wrote the researchers, Aruna Ranganathan and Xingqi Maggie Ye.

The findings are not totally unexpected. Last year, researchers from the University of Chicago and the University of Copenhagen found that AI chatbots only saved workers about an hour a week, and would actually create new tasks, to nullify any time saved.

The HBR researchers found three ways in which AI had intensified work: task expansion, blurring of boundaries between work and non-work, and more multi-tasking.

It’s another reminder that while AI is going to make the handling of routine tasks easier, its introduction comes with a range of issues. “The question facing organizations is not whether AI will change work, but whether they will actively shape that change — or let it quietly shape them,” conclude the researchers.

This article first appeared on CIO.