The disparity between the trillions spent fueling AI and the distaste of the people meant to adopt it has grown into a chasm. Only 16% of Americans believe AI will have a positive impact on society over the next 20 years, according to a recent survey conducted by Pew, while 40% expect the opposite.
There’s a number of reasons people detest AI—the data centers are disruptive, it gobbles up water—but by far the most salient one is that it could take jobs. Robert Shiller, a Nobel economist, worries that that panic could become a self-fulfilling prophecy.
In a guest essay in The New York Times on June 22 headlined “This Doommaxxing Has Got to Stop,” the Yale economist expanded on his Nobel-prize winning work on how markets misprice risk. He’s now interested in the cause of that mispricing, and the cause, he argued, is about narrative, the stories people tell each other about where the economy is headed.
“When millions of people make millions and millions of decisions based upon negative expectations, there is a risk that fear can actually help birth the reality,” he warned.
The fear that the machines are coming for workers is an old one. But in each instance, fear ran ahead of the actual displacement, according to Shiller. Luddites revolted against the loom in the 1830s, while newspapers perpetuated the drama. The 1920s got a hit play R.U.R., in which the robots rise against the people who built them.
Similarly, the 1929 stock market crash couldn’t have caused the Great Depression as only about 2% of American households owned stock at the time. What deepened the economic ruin was a collapse in consumer spending, driven by sudden, widespread uncertainty about future income.
And a 1957–58 downturn was branded the “Automation Recession” by journalists who pinned it on factory machines; it was later re-described as an ordinary cyclical dip.
Shiller worries the same misattribution is underway now. The job market has slowed for a host of reasons, he wrote, but there are reports that fear of an AI apocalypse is “worsening the freeze and contributing to record lows in consumer sentiment.”
Seventy percent of Americans told Quinnipiac in March they expect AI to leave fewer jobs for people, up from 56% a year earlier, yet the Yale Budget Lab has found no significant change in the occupational mix among the jobs most exposed to AI since ChatGPT launched in late 2022.
This round is even worse, he said, because the source of the fear is coming from the leaders of the AI buildout themselves. Anthropic’s Dario Amodei has said AI could erase half of entry-level white-collar jobs within five years; Microsoft AI chief Mustafa Suleyman has put most white-collar automation inside 12 to 18 months. Both have since walked the timelines back.
Shiller’s one note of hope is leadership; he cites a recent paper finding that President Franklin Roosevelt’s 1935 fireside chat measurably raised spending in cities with more radio exposure.
“There’s only so much Washington can do about these narratives. And, suffice to say, Donald Trump is no Franklin Roosevelt,” Shiller concluded. “As such, perhaps the best we can do is to appeal directly to the leaders of Silicon Valley who have been promoting these negative narratives with such vigor. Surely the resulting media attention highlighting how dangerously powerful your A.I. model is may help you sell more wares, but it may be far harder to do so in a period of recession.”