You'll Spit Take When You Hear How Little Time Workers Are Saving With AI, According to This Huge New Study

The use of AI chatbots has exploded in the workplace.

White-collar workers have adopted the tech in huge numbers. Executives are looking for every opportunity to shoehorn AI into every aspect of their businesses. Companies continue to pour tens of billions of dollars into building out enormous data centers to support increasingly power-hungry AI models.

But is any of this actually leading to higher efficiency at work and increased pay?

According to a recently-released working paper by a pair of economists at the University of Chicago and the University of Copenhagen that was spotted by Fortune, it's not looking great.

After analyzing data about 25,000 workers across 7,000 workspaces, users of AI only saved on average three percent of time. And only a meager three to seven percent of those productivity gains translated into bigger paychecks, they found.

Put simply, AI isn't even close to coming for everybody's jobs, a frequently cited fear — and it's not exactly making workers more productive, either, despite the industry's many reassurances.

"While adoption has been rapid, with firms now heavily invested in unlocking the technological potential, the economic impacts remain small," the two economists wrote in their paper. "Our findings challenge narratives of imminent labor market transformation due to generative AI."

As The Economist points out, while many pundits point to growing unemployment and blame AI, existing macroeconomic data suggests there are plenty of other reasons to blame. To the contrary, the magazine found that employment in white-collar work has actually risen in the past year.

"Software, writing code, writing marketing tasks, writing job posts for HR professionals — these are the tasks the AI can speed up," coauthor Anders Humlum, at UChicago, told Fortune. "But in a broader occupational survey, where AI can still be helpful, we see much smaller savings."

"I might save time drafting an email using a large language model, so I save some time there, but the important question is, what do I use that time savings for?” he added. "Is the marginal task I’m shifting my work toward a productive task?"

What workers actually do with the tiny amounts of extra free time, courtesy of AI, isn't exactly comforting, either. More than 80 percent of workers in the study said they used saved time for more work. Less than ten percent said they took the time off.

The findings "suggest that workers are not exactly knocking on the boss’s door asking for more work," Humlum told Fortune.

Signs that the impact of AI on the labor market has been overblown are certainly there. Earlier this year, Klarna CEO Sebastian Siemiatkowski admitted that his all-in approach to replacing human customer service agents with AI was not working. The company revealed last week that it's facing net losses of $99 million for the first quarter of this year, more than double compared to the same period last year.

Last week, Duolingo CEO Luis von Ahn walked back his earlier promises to replace all contract workers with AI following an immense wave of customer blowback.

Earlier this month, an IBM survey found that only a fraction of AI initiatives are delivering a return on investment after surveying 2,000 CEOs.

These latest findings are an intriguing reality check for the notion that AI is coming for all of our jobs. Perhaps the tech is both making us efficient, but also failing to make us redundant.

The conclusion highlights growing concerns that generative AI tech may be a dead end after all, as some experts have warned. Perhaps our jobs may be safe — other societal factors, like historic levels of economic uncertainty, notwithstanding.

More on AI job automation: AI Is Replacing Women's Jobs Specifically