For a long time, economists were at the forefront of advising caution against fears surrounding technological advancement. They assured us that ATMs didn’t eliminate cashiers, Excel hadn’t replaced bookkeepers, and robotic vacuums weren’t substituting maids, with ‘augment, not replace’ being their prevailing view.
However, this consensus is beginning to break down.
A recent study conducted by the Federal Reserve Bank of Chicago, along with researchers from the Forecasting Research Institute, Yale, Stanford, and the University of Pennsylvania, surveyed 69 economists, 52 AI specialists, and 38 superforecasters about the potential impact of AI on the U.S. economy.
All groups concur that accelerated progress in AI could lead to reduced labor force participation—a diplomatic way of indicating ‘fewer people working.’
The figures are striking. Under what researchers term a ‘rapid’ scenario, where AI outperforms humans across most cognitive and physical tasks by 2030, economists predict the U.S. labor force participation rate will fall from its current 62% to 54% by 2050.
Approximately half of this decline, equating to around 10 million jobs lost, would be directly due to AI, not demographics or other trends.
This ‘rapid’ scenario isn’t merely speculative. It envisions a world where AI can negotiate book contracts, assist in any factory or home, and replace freelance software engineers, paralegals, and customer service agents.
Anthropic CEO Dario Amodei has already indicated that the pace of disruption is faster than expected, aligning with the study’s rapid scenario. GDP projections also tell an important part of this story.
Under the same rapid scenario, economists predict annual GDP growth reaching 3.5% by 2045-2049—comparable to post-WWII boom levels. AI specialists are even more optimistic, forecasting a 5.3% increase. This could lead to significant wealth creation concentrated at the top, with fewer workers to share it. Researchers warn that under rapid AI development, the wealthiest 10% of households might hold 80% of total wealth by 2050—higher than pre-WWII inequality levels.
A subtlety often missed in discussions about AI and jobs is highlighted in the paper: expert disagreement isn’t primarily over whether powerful AI will arrive but what happens to the economy afterward. This marks a significant shift from earlier pro-tech arguments that assumed transformative automation would eventually create new job categories. The current debate among economists focuses on whether AI, unlike ATMs, might automate the process of inventing new tasks.
At present, overall employment data appears stable. A Yale and Brookings study from late 2025 found no significant unemployment signals nearly three years after ChatGPT’s launch. However, research cited in this paper notes a 13% relative decrease in employment among workers aged 22-25 in occupations most exposed to AI. The broader picture is stable, but the leading edge shows signs of disruption.
Regarding policy, economists and the public diverge sharply. Economists favor targeted retraining programs (71.8% support) while largely opposing job guarantees (13.7%) and universal basic income (37.4%). Conversely, the general public seems more open to structural interventions. The paper’s authors point out that optimal policy will depend heavily on which scenario unfolds—and currently, it remains uncertain.
Thus, the ‘augment, not replace’ narrative isn’t completely obsolete but is teetering on life support, with economists now having enough data to raise concerns.