Stay informed with free updates
Simply subscribe to Technology myFT Digest – delivered straight to your inbox.
The first time I heard the term “superfluous people” was when I read the 19th century Russian writers Alexander Pushkin and Ivan Turgenev. In their stories, spoiled, world-weary people of the petty nobility chased women, gambled with their inheritances, and shot each other in duels.
Like “the fifth wheel on a wagon,” as Turgenev described them, they could find little purpose in life, and their real-life counterparts would later be drawn into root causes. This elite overproduction is sometimes blamed for fueling the Bolshevik Revolution of 1917.
The second time I heard the term “redundant people” was in a recent, terrifying conversation with a West Coast venture capitalist. Only this time it was about the AI revolution. His point was that machines will soon be able to do almost all the jobs that humans currently do, making many of us unnecessary.
“There will be only two types of jobs in the future: those that tell machines what to do, and those that tell machines what to do,” he said.
In other words, you either write the algorithms that direct Uber drivers where to go. Or you'll be the Uber driver with that algorithm telling you where to go. But again, both functions may disappear with the arrival of fully self-driving cars.
This reductionist talk has become louder as the hype around AI grows. Intelligent machines will automate brain power in the same way that dumb machines automated muscle power during the Industrial Revolution. Once again, the recurring specter of technological unemployment has emerged. Billionaire businessman Elon Musk told British Prime Minister Rishi Sunak last year that artificial intelligence would be “the most destructive force in history” and we could get to a point where “jobs are no longer needed.” “AI will likely be smarter than any human next year,” Musk posted this week.
This sense of technological determinism was echoed in part at a recent Ditchley Foundation conference in Oxfordshire, UK, on the impact of artificial intelligence on work and education, which was attended by policymakers, technologists and business managers. Some speakers said we were rapidly approaching a jobs “emergency.” Employers were already exploiting the potential of generative AI to shed workers and reduce graduate hiring. Today, generative AI threatens the jobs of copywriters and call center workers. Tomorrow it will strike middle managers and lawyers.
Generative AI will also change the nature of many tasks employees perform, even if it does not kill their jobs completely. One study of its impact estimated that technology will affect at least 10 percent of the tasks performed by about 80 percent of the U.S. workforce.
But some labor market experts consider these sweeping predictions of a jobs apocalypse to be ahistorical and almost certainly wrong. They ignore our past experience with new technologies, the dynamics of societal adaptation, the possibilities for creative innovation, and the weight of demographics. In short, they confuse technological feasibility with economic feasibility, as sociologist Aaron Benanav has argued.
One of the main complaints of employers at the Ditchley conference was how difficult it is to recruit skilled workers in economies near full employment and societies with aging populations. While it is easy to see the jobs that AI will disrupt, it is difficult to imagine the jobs that will be created. About 60 percent of job categories in late 2010 did not exist in 1940 — in medicine, software, entertainment, and solar energy, for example. “Unless there is a massive change in immigration policy, the United States and other rich countries will run out of workers before we exhaust ours,” David Autor, an economist at the Massachusetts Institute of Technology, wrote in a recent article.
As Autor and others have argued, we should therefore view AI as an opportunity, not an emergency. It provides the opportunity to expand the “importance, extent and value” of human experience to include more workers and rebuild the middle class.
We can use AI to enhance lifelong learning and supplement a dwindling workforce. We can upskill and re-evaluate professions that humans still do best, such as nursing and teaching. We must find better ways to redistribute the financial gains of the AI revolution from winners to losers.
Failure to do so may lead to another revolt by the “superfluous,” only this time against the robots, not the Romanovs.
john.thornhill@ft.com
