🤖👨‍🏭 The AI Revolution: Still no reason to fear a ‘jobpocalypse’

History and economic data give reasons for optimism

Quote of the Issue

“My guess is that we’ll have AI that is smarter than any one human probably around the end of next year,” said the billionaire entrepreneur, who runs Tesla, X and SpaceX. Within the next five years, the capabilities of AI will probably exceed that of all humans, Musk predicted on Monday during an interview on X with Nicolai Tangen, the chief executive of Norges Bank Investment Management. - Financial Times, 04/08/2024

---

The Conservative Futurist: How To Create the Sci-Fi World We Were Promised

“With groundbreaking ideas and sharp analysis, Pethokoukis provides a detailed roadmap to a fantastic future filled with incredible progress and prosperity that is both optimistic and realistic.”

---

The Essay

🤖👨‍🏭 The AI Revolution: Still no reason to fear a ‘jobpocalypse’

The 2010s saw the biggest resurgence in fears about technological unemployment since probably the 1960s (a phenomenon I outline in The Conservative Futurist). The big macro explanation, I think, concerned a) the slow decline in unemployment after the Global Financial Crisis, and b) the emergence of the Digital Revolution’s next phase, including app-enabled smartphones and advances in machine learning, such as data mining and machine vision. It was certainly tempting to connect the dots and think robots were beginning to take all the jobs. 

By 2020, for instance, dark-horse candidate Andrew Yang was able to make a splash in the Democratic presidential primaries by calling for a universal basic income due to the impending tsunami of job-killing automation. As Yang told the New York Times in 2018,

All you need is self-driving cars to destabilize society. [In just a few years] we’re going to have a million truck drivers out of work who are 94 percent male, with an average level of education of high school or one year of college. That one innovation will be enough to create riots in the street. And we’re about to do the same thing to retail workers, call center workers, fast-food workers, insurance companies, accounting firms.

Yang’s worries about autonomous driving didn’t come from nowhere. Plenty of folks in Silicon Valley, including Elon Musk, were claiming self-driving cars were nearly ready to hit the road en masse. What’s more, the subject of technological unemployment was getting serious treatment from economists.

Living in the back half of the chessboard

In their highly (and still) influential 2011 book, “Race Against the Machine,” Erik Brynjolfsson and Andrew McAfee of MIT explained that the accelerating digital economy meant “the threat of technological unemployment is real. ... As the technology moves into the second half of the chessboard, each successive doubling in power will increase the number of applications where it can affect work and employment. As a result, our skills and institutions will have to work harder and harder to keep up lest more and more of the labor force faces technological unemployment.”

It’s a conclusion that probably really struck home with readers among the millions of Americans that year who watched IBM's Watson AI system defeat two of quiz show Jeopardy!’s all-time human champions.

Other economists attempted to measure worker vulnerability to smart machines, including artificial intelligence. Perhaps the most famous estimate came from a 2013 analysis by Oxford University researchers Carl Benedikt Frey and Michael Osborne who claimed that 47 percent of total US employment was in the “high risk category, meaning that associated occupations are potentially automatable over some unspecified number of years, perhaps a decade or two.”

To arrive at this estimate, Frey and Osborne, along with a group of ML researchers, subjectively hand-labeled 70 occupations, assigning 1 if automatable, and 0 if not. They then created a computer program that learned from this labeled data and the key job traits that make automation harder. After training on those jobs, the program was used to calculate the probability that computers could do each of some 700, based on their automation traits. 

(Breaking jobs down into their specific tasks and then examining them for their automation potential isn’t dissimilar to the method followed by a recent Goldman Sachs analysis that finds most jobs and industries are only partially exposed to automation, and are thus more likely to be complemented, rather than substituted, by generative AI “if it delivers on its promised capabilities.” The bank’s economics team finds that “7 percent of current US employment could be substituted by AI, 63 percent could be complemented, and 30 percent could be unaffected.”)

Furthermore, Frey and Osborne suggested that many jobs in transportation, logistics, office support, and production were at high risk of being automated. Many service jobs, a major source of US job growth for decades, were also vulnerable. Additionally, they determined that lower wages and education levels were strongly linked to a higher likelihood of a job being automated. That would represent a shift from the 20th-century trend of middle-income jobs being most affected. To stay competitive, they concluded, “low-skill workers will reallocate to tasks that are non-susceptible to computerisation—i.e., tasks requiring creative and social intelligence.”

So far, no jobs collapse in sight

Back to the present: The robots and algorithms have yet to cause mass technological unemployment, much less take all the jobs. The jobless rate is at decades lows. There’s such a need for workers that one explanation for the continued high rate of US job growth is the surge in immigration, with roughly 2.5 million newcomers arriving in 2023, the highest level in the last two decades. (Goldman Sachs notes that “recent immigrants have a higher labor force participation rate (64%) than the native-born population.”)

Still, the emergence of GenAI has reenergized tech unemployment concerns, a fact noted in a new analysis from Frey and Osborne. “Generative AI and the Future of Work: A Reappraisal” updates their 2013 paper.