Why Replacing Workers With AI May Be The Biggest Mistake Leaders Can Make

This article by Kit Eaton was originally published on Inc.com.

AI is possibly the most hyped technology revolution of all time. But it’s so controversial that there are three competing arguments about its impact. AI boosters insist it’s about “augmenting” workers to boost productivity. AI critics worry that it’ll replace workers as cutthroat leaders seek profits. And AI doubters suggest that despite the hype, the tech isn’t delivering benefits at all. Now new research adds some clarity to the debate, and it’s actually good news for flesh and blood workers: it says adding AI to a business can lift productivity, and even boost wages. But only if it’s used as a “helper,” not a replacement.

Expert economics and tech researcher Christos Makridis, writing at The Conversation, explains that his recently published paper looked at how AI impacted industries across the U.S., using data from 2017 to 2024 covering “nearly all employers” during the period before, during, and immediately after tools like ChatGPT arrived. 

The data showed that industries that adopted AI earlier, because it fit their needs, tended to be using automation tech/social media tools and were benefitting from productivity growth even before AI arrived. But strikingly, for data from 2024, the industries that had more AI “exposure” (i.e. AI tools fitted their needs well, and were embraced fast—like the software development industry) saw gains of up to 10 percent productivity compared to industries that didn’t embrace AI in the same way. Amazingly, these industries also saw up to four percent rises in jobs and nearly five percent rises in wages. 

The patterns imply, Makridis explained, that “at least so far, AI has acted as a productivity-enhancing tool that boosts employment and wages rather than a simple substitute for labor.” 

This is a fascinating result, and it’s one in the eye for AI critics. But the research went deeper and found bigger implications for the whole world’s job markets.

There is a subtlety in Makridis’ data, connected to the difference between jobs where AI works with people, and where it can be deployed more autonomously. In some industries where AI is complementary (Makridis suggests “marketing, writing or financial analysis,”) the data showed employment rose as AI exposure rose. But in industries where AI can run more autonomously, like basic data analysis, there was no big employment change, even if workers saw comparatively slower wage growth.

Thus, overall, Makridis argues that his research shows that, on average, job markets are “more resilient than the most alarmist forecasts suggest.” 

What it all boils down to is the utility and value of deploying AI. When AI helps drop the cost of finishing a task, and boosts worker productivity, savvy employers benefit and can lift their output … which then increases their need for workers.

This is identical logic to that which explains “why power tools didn’t eliminate construction workers,” Makridis argues: power tools boosted worker productivity, but didn’t replace workers, and lifted the whole industry. New tech often impacts the world of work, but it also “rarely” eliminates the need for human workers—instead, it just changes what work happens. “Some tasks shrink. Others expand. New ones emerge that were previously too costly or too hard to perform at scale,” Makridis says.

What needs to happen is that in the sparkling, exciting light of AI, companies and leaders must “create environments where workers can experiment, reorganize tasks and integrate new tools into productive routines,” he suggests. Making workers feel like they have the “psychological safety to experiment” means they may use AI more and in more productive ways, boosting higher-value work, he says. 

What does all this mean for you and your own company’s embrace of controversial AI tech?

First up, it implies that if you’re planning on outright replacing some of your staff with AI, you may be making exactly the wrong move.

If you’re tech-savvy and sensitive to your employees’ worries about the threat of AI, then you can quote Makridis’ research as an example of how AI is better at boosting workers rather than replacing them. A hearty “we’re all in this together” discussion, centering around the best practices of AI use and how it’ll benefit everyone, including your staff, could go a long way to assuage workers’ tech fears. You can even point to how the data suggests moves like Jack Dorsey’s firing of 4,000 workers from his fintech company Block in order to embrace more AI may ultimately prove a failure.

The next step then is to encourage an open and free attitude to using AI—your workers may find benefits, business-aiding changes, and quick fixes from using the tool that had never occurred to you. Just remind your staff to always check and verify the output of an AI tool: as yet, it’s not as reliably truthful or accurate as they are.