Business
‘Humans could go the way of horses’: Goldman calculated how bad the AI ‘job apocalypse’ will be
In 1983, the Nobel Prize-winning economist Wassily Leontief asked whether technological change could become so profound that “humans could go the way of horses” when tractors replaced them in agriculture and transport in the early part of the 20th Century.* Might not computers replace the need for humans who can think the same way the combustion engine replaced the need for literal horsepower?
This week, two analysts at Goldman Sachs tried to answer that question in a research paper cheerfully titled, “How Concerned Should We Be About a Job Apocalypse?”
Quite, but not too much, is their conclusion.
Joseph Briggs and Sarah Dong estimate, based on Department of Labor job numbers, that 25% of all work hours could be automated by AI. Thus, “We expect that the AI transition will lead to a meaningful amount of labor displacement.”
AI won’t replace jobs in a uniform way, however. “Our baseline forecast for a 15% AI-driven labor productivity uplift and the historical relationship between technologically driven productivity gains and job loss implies that 6-7% of jobs will be displaced over the adoption period,” they said.
“We estimate a peak gross unemployment rate increase of around 0.6pp (corresponding to a 1 million increase in unemployed workers.”
That sounds bad, but there is good news.
Previous eras of technological change have resulted in the creation of a mass of new jobs that no one previously was able to imagine, the Goldman team said.
“Technological change is a main driver of long-run job growth via the creation of new occupations—only 40% of workers today are employed in occupations that existed 85 years ago—suggesting that AI will create new roles even as it renders others obsolete.”
“More than 6 million workers are currently employed in computer-related occupations that did not exist 30-40 years ago, while another 8-9 million are employed in roles enabled by the gig economy, e-commerce, content creation, or video games.”
Fundstrat head of research Tom Lee recently made a similar comparison in an appearance on the Prof G Markets podcast with Scott Galloway and Ed Elson, comparing the current AI boom to the introduction of flash-frozen foods in the 1920s. Citing his firm’s research, he claimed this reduced farm labor from 40% of the U.S. workforce to 2%, but enough new jobs were created that the shift was overall positive.
“Let’s say there was a CNBC in 1920 and these economists were saying, ‘frozen food, if it comes along and it’s going to wipe out 95% of all farmers, this is going to wipe out the U.S. economy. The U.S. economy can’t survive frozen food … Instead it freed up time, right? And it created, it allowed people to be repurposed, and it created a completely new labor force.”
*Leontief originally wrote, “The role of humans as the most important factor of production is bound to diminish in the same way that the role of horses … was first diminished and then eliminated.” This has been truncated over time and is now widely attributed to him as, “Humans could go the way of horses.”