By Peter Cappelli 

How fast is AI advancing in the workplace? 

Apparently, not very.  

I was sitting in a meeting of top executives when I was struck by the fact that there seemed to be a uniform belief that AI and its associated tools were rolling through companies at real speed, cutting jobs left and right. 

There is still no evidence for this view. 

I am not good at making predictions, including ones about AI. My interest is in seeing what is actually happening with the introduction of AI in the workplace. It has now been four years since ChatGPT burst on the scene introducing us to large language models (LLMs).  Longer than you thought, right? My colleagues and I have been studying it since the late 2010s, so let’s see where it stands now. 

The first issue is how widespread its use is. The Census reports that less than 20% of U.S. businesses are using it at least once in a two-week period (surveyed in May 2026), a figure that has held steady over the past six months. In other words, not increasing at a break-neck pace, and it is much lower than I would have expected, especially given that most of us probably use it every time we use a search engine.    

Turning to individual use at work, is it really taking over much of the daily tasks of typical workers? It is possible, in fact likely, that employees are using it without the knowledge of their employers. That is not a good thing if, for example, decisions are being made based on LLM outcomes that haven’t been checked. Gallup data says that 28% of employees use it at least a few times per week. But are they using it to pick lunch spots or something more fundamental to business?   

To get at that question and how work is performed and whether AI is changing it, turn to a recent survey by Enhancv.com, a career site. It asked respondents whether AI tools defined broadly had reduced their time on traditional tasks by 50% or more. That would be enough to suggest that we could see job cuts. Only 5.6% said yes; 27% said it had some effect but not that much; 36% said it had no noticeable impact, and 23% reported that it wasn’t applicable to their job. In other words, nada for 59%. 

Two big questions come from these data points. First, why is it that business leaders feel that AI is sweeping across the economy?  One reason might simply be the hype cycle: The initial claims about what AI would do were apocalyptic with all low-skill jobs being wiped out, no entry-level, white-collar jobs surviving, and so forth. But no headlines in the business press are going to scream: “Nothing happening here!”    

There has been some pushback from the press, especially against the idea that the current round of layoffs has been caused by AI. Surprising findings from a survey of CFOs indicate that only 44% thought that AI would cut jobs. That group reported only a trivial 0.14% reduction in jobs in 2025 was due to AI.  The majority reported no cuts due to AI. But it takes a while for perceptions to change and especially to undo initial impressions.   

But why wouldn’t leaders have a better sense of what was happening in their own companies? Maybe they aren’t asking, maybe they are being told what they want to hear.  

The second question: Given how easy it is to use these AI tools, why is it taking them so long to have an impact? It is not a skills problem. The idea that you need training to learn how to use ChatGPT is a joke. That is not what is holding up its use.   

To be clear, any tool that increases productivity is good and necessary. It is what makes us better off. It is, for example, why the US economy is four times bigger than it was after WWII while the population only doubled. We can each get twice as much done now because ways were found to improve productivity.    

What appears to be holding it up are issues of organizational change. It takes a lot of management attention to introduce these AI tools. Now that CFOs no longer think AI is an easy way to cut headcount, their checkbooks are tightening up. Maybe that’s a smart thing for businesses. Expect a slower and probably more sensible advance.   

Peter Cappelli is the George W. Taylor Professor of Management and Director of the Center for Human Resources for The Wharton School.