Is AI really enabling productivity gains?

More work in less time with fewer workers. Productivity gains are supposed to be one of the big benefits of artificial intelligence. But a new survey of corporate executives around the world suggests those promises have not yet come to fruition.

More than 80% of the 6,000 executives surveyed by the National Bureau of Economic Research “detect no discernible impact from AI on either employment or productivity,” said The Register. It is not for lack of trying: 69% of businesses say they use AI in the workplace and three-quarters “expect to use it over the next three years.” More than 90% say artificial intelligence had “no impact on employment” at their businesses. The new survey is the latest addition to a “growing body of evidence” that AI’s advocates are “just not living up to their promises — at least not yet.”

The link between AI and productivity is “murky at best,” said Marketplace. That is because any productivity improvements are “going to be really hard to measure,” said Erika McEntarfer of the Stanford Institute for Economic Policy Research to the outlet. There are other factors increasing business productivity at the moment, including new investments in research and the “loosening labor market,” said Marketplace. Figuring out AI’s impact will involve measuring “hundreds of millions of people, doing at least that many, if not more, discrete tasks every day,” said George Pearkes of Bespoke Investment Group.

The Week

Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

SUBSCRIBE & SAVE

Sign up for The Week's Free Newsletters

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

Sign up

What did the commentators say?

The NBER survey is “damning,” said Frank Landymore at Futurism. While most firms are using AI in some fashion, the “vast majority” say the technology “hasn’t budged the needle for them yet.” No surprise. Other surveys have found that AI can “slow down rather than speed up human programmers” and ends up “accelerating burn-out” among human workers. There is precedent for this: The adoption of computers decades ago was “obviously transformative,” but they “didn’t immediately translate to economic gains.” Which is why executives will keep “clinging to the hope that the tech’s promises will be borne out in the long run.”

Businesses are experiencing the “pause before the gale,” said James Pethokoukis at the American Enterprise Institute. There is a growing consensus that AI will gradually seep into the workplaces via office software in “useful, but hardly revolutionary” fashion. The firms that see productivity gains will be willing to “thoroughly rethink how work is organized.” When the promised benefits of AI finally arrive, “no one will doubt its existence and import.”

What next?

AI’s economic impact is “just beginning,” said Columbia Business School. But the gap between the promises and the measurable outputs is creating a “growing tension in public discourse.” Artificial intelligence already “feels transformative” in many users’ daily lives, but the “effects are not fully visible in traditional macroeconomic statistics.” What seems certain is that work will evolve as the technology changes. Workers have adapted to new technologies throughout history, said Aaron “Ronnie” Chatterji, OpenAI’s chief economist. “I’m bullish on humans,” he said.