The K-shaped economy has dominated recently, but in the debate over the impact of artificial intelligence on productivity, the J-shaped economy is also starting to enter the discussion.
This curve refers to general-purpose technologies such as artificial intelligence that do not yield immediate benefits. Instead, massive investments come first, masking early gains. Only after the initial dip does productivity really take off, creating a J shape. But for some, it’s unclear whether that shift is happening.
Torsten Slok, chief economist at Apollo, quipped: “Beyond incoming macroeconomic data, AI is everywhereRecall Robert Solow’s famous quote about the personal computer revolution. Slock added that employment, productivity and inflation statistics still show no signs of new technology. Meanwhile, margin and earnings forecasts for S&P 500 companies outside the “Magnificent 7” also lack evidence of AI’s role.
“Maybe there is a J-curve effect in AI, and it takes time for AI to show up in macro data. Maybe not,” he wrote in an article. Watch out for Saturday.
but in a financial times Column Economist Erik Brynjolfsson, titled “The Takeoff of Artificial Intelligence Productivity Is Finally Visible,” pointed to the latest U.S. Bureau of Labor Statistics jobs report as evidence that “the fog may finally be lifting.”
Wednesday’s report revised the 2025 job growth figure to just 181,000 jobs, down from the 584,000 initially reported and down from 1.46 million job gains in 2024.
GDP grew 3.7% in the fourth quarter, pointing to a sharp rise in productivity, given that the economy continues to expand at a healthy pace and added few workers last year.
Brynjolfsson said his own analysis showed that U.S. productivity would grow by about 2.7% in 2025, nearly double the 1.4% average annual growth rate over the past decade.
“The latest U.S. data for 2025 suggest that we are now transitioning from the investment phase to the harvest phase, with early efforts starting to show up as measurable output,” he said.
Brynjolfsson, director of the Digital Economy Laboratory at Stanford University, who has been studying artificial intelligence before ChatGPT shocked the world, published an article First-of-its-kind research Last year showed that artificial intelligence Disproportionately hitting entry-level workersespecially those aged 22 to 25 who are engaged in highly artificial intelligence occupations.
He warned that several more episodes of sustained growth would be needed to confirm long-term trends in productivity, adding that geopolitical or currency dislocations could offset progress.
But while many businesses are still using AI in a minimal way, Brynjolfsson said he’s seeing “a small group of power users” using AI agents to automate end-to-end workflows, completing tasks in hours instead of weeks.
“We are transitioning from an era of artificial intelligence experimentation to an era of structural practicality,” he writes in Artificial Intelligence. financial times. “We must now focus on understanding its precise mechanisms. The recovery in productivity is more than just an indicator of the power of AI. It is a wake-up call to focus on the coming economic transformation.”
Others also see clear signs that AI is improving productivity when looking at the information and communications technology (ICT) industry.
Stephen Brown, chief deputy economist for North America at Capital Economics, said in a report earlier this month that ICT output grew in the third quarter despite a decline in employment.
He added that while earlier layoffs may have been due to over-hiring during the pandemic, layoffs are continuing even as the ICT industry is booming.
“All of this means that AI is making a huge contribution to productivity growth,” Brown declared.