When CEOs talk to investors about layoffs, they usually blame economic uncertainty or business “headwinds.” Now a new term is starting to crop up in these announcements: AI.
In recent months, shipping giant UPS announced plans to cut 12,000 office jobs that CEO Carol Tomé said were unlikely to return because the company was increasingly using AI to automate tasks these workers performed. Meanwhile, financial giant BlackRock said it would eliminate about 600 positions, couching the cuts as an effort to prepare for coming shifts in the asset management industry, of which AI is among several drivers. Then there was Google, which recently laid off ad-sales staff partly because new AI tools were helping customers manage ad campaigns themselves. And IBM CEO Arvind Krishna said Big Blue would pause hiring for 7,800 roles because AI could now do the work instead.
To some commentators, these developments are early confirmation that long-standing predictions of AI causing widespread job loss, perhaps even ushering in an era of “mass unemployment,” are about to be realized. This narrative of AI-driven job loss is so ingrained in the public consciousness that people often see evidence for it when none exists.
While the U.S. unemployment rate rose slightly to 3.9% in February, it remains near historic lows. Yet when executive search firm Challenger Gray & Christmas released a report saying there had been 4,600 U.S. job cuts directly attributed to AI from May to January, some commentators interpreted the relatively low number as evidence that firms were hiding the true extent of AI casualties because they feared public backlash.
It’s possible that one day there will be a significant number of jobs lost because of AI. Goldman Sachs forecast AI software could automate the equivalent of 300 million full-time roles globally by 2030. But that day has not yet arrived, economists and business analysts say. “This is not that moment,” notes Daniel Susskind, an economics professor at King’s College London and the author of several books on technology’s impact on work, including the forthcoming Growth: A History and a Reckoning.
Right now, wide-scale AI deployment is held back by several factors. Many companies are still just trying to figure out how to use it in ways that would justify the sky-high cost. Furthermore, Boston Consulting Group recently reported that most top executives are anxious about using AI across their organizations given the risk of “hallucinations,” where AI software produces sometimes dangerously inaccurate information.
So, why then are some CEOs mentioning AI in the same breath as layoffs? Some of it is just marketing spin, experts say. Erin Ling, an assistant professor specializing in AI and the future of work at the University of Surrey, in the U.K., said AI is often simply a convenient cover for layoffs that are the result of bad management, struggling businesses, and deteriorating economic conditions. For public companies, the bad news that they’ve had to cut jobs because of financial distress “becomes slightly better news because of AI,” says Grace Lordan, an economist and professor of behavioral science at the London School of Economics. “It looks like smart cost cutting.”
This certainly seems to be the case with UPS. At the same time the company announced layoffs and mentioned its increasing use of AI, it delivered the unpleasant news to Wall Street that it had missed forecast revenue and earnings numbers and was lowering its revenue guidance for the coming year. Package-delivery volumes are also down. AI was pretty much the only positive thing Tomé mentioned.
300 million
Number of full-time jobs that will be fully automated globally by 2030. Source: Goldman Sachs
The situation at Google and some other tech companies is more nuanced. Here they aren’t simply trying to burnish their tech street cred. They already have that in spades. Instead, the layoffs are about cutting costs to invest more in developing AI, because the computing resources and human machine-learning talent needed are so pricey. So in these cases, AI is, in fact, linked to job loss—but not for the reasons people have long feared.
Carl Benedikt Frey, the University of Oxford economist who coauthored one of the first landmark studies of AI’s potential impact on jobs, says that people are probably overestimating the job loss generative AI will cause in the near term. “Generative AI is not an outright automation technology,” he says, noting that people are still needed to write the prompts that are fed to the software, and to check the quality of its output. “You need a human in the loop in most cases.”
He’s among those who think we could see a significant “Uber effect” from AI, where the technology lets less skilled and less experienced workers take on higher-level tasks. Uber allowed anyone with a driver’s license and a car to potentially become a taxi driver. As a result, many more people became drivers for hire.
Similarly, AI “copilots” could help many more individuals perform legal, financial, or software-coding tasks. Rather than eliminating jobs in these fields, this technology could help their ranks soar, Frey says. That’s because all evidence suggests there is a large demand for professional services that is not currently being met, partly because such services are too expensive for many customers to afford.
But just as Uber was bad news for taxi drivers, who struggled in the face of lower-cost competition, certain existing employees may see their wages fall—or at least stagnate—because of AI. On the other hand, even these reduced wages might be more than what less-skilled workers can earn in other fields today. So, overall, economic inequality could be reduced.
Frey is less sanguine about AI’s longer-term effects, however. He says AI may currently be in its “lamplighter” phase: When streetlights were powered by gas, people were employed to ignite each lamp every day at dusk using a flaming wick carried on a long pole. When electric bulbs were introduced, the lamplighters kept their jobs because each streetlight had to be switched on individually. But soon cities began installing switches that controlled entire city blocks, and eventually timers and light sensors meant no human intervention was needed at all. Frey thinks AI could well follow a similar path, with today’s period of relatively little job displacement lulling us into a false sense of security.
Almost everyone agrees AI is ushering in an era of uncertainty and disruption and that workers will need to be prepared to learn new skills and shift roles. Many experts argue governments should do more to encourage lifelong education and retraining. And Susskind says governments should eliminate tax incentives that encourage businesses to use AI to replace workers rather than augment them.
Taking these steps now could mean that the AI-doomsday scenarios of mass unemployment never come to pass. At the very least, we should stop panicking about CEO pronouncements on AI and layoffs—and just get back to work.
This article appears in the April/May 2024 issue of Fortune.
