This Labor Day, instead of celebrating work, many Americans may be worrying about how much longer they will even have jobs in the age of artificial intelligence.

With critics warning millions of workers will be replaced by robots and computer programs, it is unsurprising many workers worry they’re about to be tossed aside.

It is just as likely that AI will lead to employment gains.

A recent study, apparently the first of its kind, by the financial technology firm Ramp looked at company payments to AI vendors and compared that to the payrolls for the same companies. It found that AI adoption was strongly associated with a 10% increase in employment at companies that invested heavily in it. That’s because the more AI tech a company has, the more people it needs skilled in using that technology.

That’s not to say AI won’t disrupt the economy. Jobs will be shifted around, and many workers will have to learn new skills to stay competitive. Some companies like Microsoft have laid off some workers and redirected the savings from that into AI research. But that is not the same thing as using AI to replace workers, and it has likely created jobs at the AI companies.

Technological innovation has historically created new, better jobs, as I note in a recent study of AI and jobs for the Competitive Enterprise Institute.

Earlier fears about technology stealing jobs focused on devices, usually in factories, that replaced manual labor. What is different about AI is that it supposedly threatens white collar jobs like journalism, graphic design, and legal research. Nobody, by contrast, worried about the Internet taking jobs away. As economist Bill Beach, former head of the Bureau of Labor Statistics, told me, AI is not only a new technology, “but it is also a new form of labor, and that is the part that is puzzling everyone.”

The notion that AI will lead to thinking-robots that replace workers remains science fiction. AI is a fundamentally backwards-looking technology: It takes information and synthesizes it. It can guess at what information a user wants (i.e., run the odds based on the data it has), but the output remains guesses, and the responses can be spectacularly wrong. Teaching AI the difference between fact and fiction isn’t easy. It can “hallucinate” and literally make stuff up.

In short, AI requires much handling to use effectively. The Ramp study noted that job growth at AI-using companies was greatest at those that used AI the most extensively. The hiring effect was negligible at companies that didn’t invest in AI as much. This is in keeping with the history of new technology and jobs. The benefit new technology brings is greater production. It requires workers with different skill sets, but not necessarily fewer workers. Companies that try to simply replace workers with AI programs will likely produce shoddy products and get their comeuppance from consumers.  

When Detroit began automating assembly lines in the 1980s, that meant fewer people on those lines because robots did the most repetitive tasks. Many workers got other jobs in the same industry. Auto companies still needed people to build, maintain, and manage the robots. And the more cars rolling off the assembly lines meant that, for example, more people were needed to work the logistics of getting the cars into show rooms and, ultimately, on the roads. Employment by the Big Three auto companies has remained at about a million for decades.

AI has admittedly developed a bit of bad rap, especially lately. Tech gurus have vacillated between claiming it will create vast wealth for its users to warning it will eliminate entire sectors of workers. The construction of the data centers needed for AI’s broad adoption has become a political hot potato, driven by fears that the centers will drive up utility costs, among other concerns. Some companies, from health care providers to creative industries, are wary of AI technology when they don’t know what regulations are potentially coming and if a popular backlash will make adopting it a public relations problem.

History teaches us that worries and problems with new technologies will resolve over time. In the meantime, let’s let investors risk their own money, keep taxpayer dollars out of it, and keep the regulations light until we have a better understanding of the technology.

As for workers, they will still be needed — even as the work itself changes. So go ahead and enjoy that backyard barbecue over Labor Day weekend. There will still be jobs worth taking a day off from in the future.

BY SEAN HIGGINS | Competitive Enterprise Institute

Sean Higgins is a research fellow at Competitive Enterprise Institute and author.