But are they using it themselves? Not as much as you might expect. Some CEOs and other corporate executives are spending less time with the technology than their employees, and more than a quarter of them aren’t using it at all, according to new data. 

The gap risks deepening the divide between workers and leaders over how—and how much—AI will boost productivity and alter future employment.

On the carrot side, bosses are incentivizing workers to adopt and experiment with AI by handing out cash bonuses and awarding merch to those who comply. More stick-like tactics to spur AI adoption include tracking workers’ AI usage and factoring AI fluency and enthusiasm for the technology into performance reviews. 

But behind the doors of the C-suite, the CEOs and executive teams imposing those rules are largely only casual users of AI themselves. Nearly 70% of CEOs, CFOs and senior executives are using AI at work less than an hour a week—including 28% who never use it—according to a new survey of more than 6,000 senior executives across four countries (the U.S., the U.K., Germany, and Australia) co-authored by renowned Stanford economist Nicholas Bloom and 12 other scholars. Some leaders reported more frequent AI use: 24% of respondents report one to five hours of AI use per week, while 7% report using AI more than five hours in a typical work week. 

In the U.S., bosses are using AI 1.7 hours per work week, slightly less than their employees, who are using it 1.8 hours per week, according to the authors’ survey of 3,000 U.S. employees. (Those surveyed do not necessarily work at the same firms as the executives.) 

What’s maybe even more noteworthy is the yawning gap between workers’ and executives’ expectations around AI. When executives responded to the survey, they said they predict AI will reduce employment by 0.7% in all firms and 1.2% in U.S. firms. Contrast that to employees, who predict AI will increase employment by approximately 0.5% in their firms over the next three years. Executives also are far more bullish than employees on the productivity gains that AI will bring, predicting 2.3% growth in the U.S. over the next three years versus workers’ 0.9% forecast. 

The new data reinforces the disconnect already present between C-suite leaders who hype AI and the real-world experience of frontline users of the technology. Study after study has found that use of AI in the workplace is causing workload creep and cognitive overload. Even if it does create efficiencies, employees tend to immediately fill the time saved with more work. That may sound like a dream for business leaders, but it’s creating a workforce that’s burned out to the point of making poor decisions. 

Bloom acknowledges that there are “two different stories here.” On the one hand, “employees know best. Their bosses have drunk the Kool Aid,” Bloom said in a talk explaining his research at Charter’s Leading with AI Summit in San Francisco earlier this month. “And employees are realistic.” But there’s another, perhaps more chilling, argument that Bloom says he buys a bit more: “Employees probably don’t see the full picture.”

He recounted a recent case of an executive who’d heard his employees talking about how AI might make their work more efficient. But the executive said his own thoughts were “Am I just going to get rid of that employee or reshape that entire division?’”

“I suspect execs are probably more correct,” he said, “but there is a clear misalignment here across society.”