Examining AI users’ digital activity 180 days before and after they began using such tools on the job, ActivTrak found AI intensified activity across nearly every category: The time they spent on email, messaging and chat apps more than doubled, while their use of business-management tools, such as human-resources or accounting software, rose 94%.

Meanwhile, the amount of time AI users devoted to focused, uninterrupted work—the kind of concentration often required for figuring out complex problems, writing formulas, creating and strategizing—fell 9%, compared with nearly no change for nonusers.

“It’s not that AI doesn’t create efficiency,” said Gabriela Mauch, ActivTrak’s chief customer officer and head of its productivity lab. “It’s that the capacity it frees up immediately gets repurposed into doing other work, and that’s where the creep is likely to happen.”

Such habits aren’t exactly what AI evangelists have predicted. A number of tech and business leaders, from Bill Gates to JPMorgan Chase’s Jamie Dimon have suggested that AI could ultimately lead people to work less, not more, and result in a shorter workweek. Elon Musk has said that, within 20 years, advancements in AI and robots could even make work optional.

Yet, evidence so far suggests that many AI adopters aren’t using the technology’s efficiencies to give themselves a break.

Dean Halonen, co-founder and chief revenue officer of software startup Steelhead Technologies, said he has experienced the work-creep first hand. Deploying AI has let his company automate a lot of administrative tasks and made its software developers more efficient at writing code, he said.

“But what we’re finding is, the work that is out there, it seems unbounded,” he said. “It’s like the appetite is always to do more, not to, like, go home at noon.”

The ActivTrak analysis backs up findings of an eight-month study on how generative AIs is shaping work habits at a tech company with about 200 employees. The research, still under way, has so far found the tools didn’t reduce work, but intensified it. The employees worked at faster paces, took on broader scopes of tasks and ended up working more hours.

People often end up doing more work, not less, “because AI makes additional tasks feel easy and accessible, creating a sense of momentum,” said Aruna Ranganathan, associate professor at the University of California, Berkeley’s Haas School of Business, who is leading that study. The initial findings were published in a recent Harvard Business Review article.

Such shifts in behavior may boost productivity but they should also be a warning sign to employers, Ranganathan says. “Over time this can lead to cognitive overload, burnout, poorer decision-making, and declining work quality, even if workers appear more productive in the short run.”

Maneesh Anand, who leads an engineering team at a telehealth startup, said the AI agents the team works with not only enable them to perform multiple tasks at the same time. They also often prompt them to dig deeper on existing projects.

“They’ll ask you, ‘Do you want me to consider this? Do you want me to consider that?’” he said. “I’ll build an implementation plan, and they’ll layer on five or six things that either you didn’t think about, or that weren’t part of the requirement.”

The ActivTrak analysis found that AI adoption is growing quickly at work, even if many workers say it isn’t saving them much time so far. About 80% of employees now use AI tools at work—up from 53% two years ago—while the average time spent working with AI tools has risen eightfold, ActivTrak said.

People who spent 7% to 10% of their total work hours with AI tools showed the highest productivity, yet only 3% of AI users use such tools that much. The majority spent 1% of their total work hours using AI.