While many Americans shudder at the prospect of AI taking their jobs, business leaders and tech enthusiasts continue praising its potential, an optimism that is echoed across Silicon Valley and Wall Street. But all that hype may actually be injuring the economy in the short term.
In a blog post from the St. Louis Federal Reserve Bank, economists argue that AI optimism could hinder productivity and act as a news shock that shapes household and business decision-making. The authors, Fed economists Miguel Faria-e-Castro and Serdar Ozkan, explain that when households see a news shock like AI adoption, they interpret it as a sign of a future pay raise, spending more today on the assumption that more money will come down the line. The same logic holds true for businesses: If you were to buy into the promise of miracle innovation—cutting the cost of labor and boosting productivity—you’d increase investment in that product. All of that enthusiasm leads to inflation in the short term as demand outpaces supply.
“Together, these forces produce an inflationary surge in aggregate demand—the defining feature of the news shock’s initial phase,” the post’s authors wrote.
AI hype is everywhere. It’s in tech entrepreneur Matt Shumer’s viral post in February, comparing the current trajectory of AI development to the month before the COVID pandemic upended the globe. It’s in the words and minds of tech leaders, from Elon Musk to Dario Amodei to Mustafa Suleyman. The technology is now creeping into the lives of workers at law firms, startups, and consultants.
Anticipated productivity gains and the dot-com bubble
While consumer prices have stabilized from a high of about 9% in June of 2022, inflation remains stubbornly above pre-pandemic levels. The most recent, the consumer price index rose 0.3% from the previous month, rising 2.4% from a year ago. While it’s hard to tell if AI hype is having an impact on prices, the researchers argue the technology could be driving up prices today. But they caution that their assessment is merely qualitative: they can predict inflation could rise in the short term, but they can’t exactly predict by how much.
The economists compare the AI hype to the optimism surrounding dot-com technology at the turn of the century. “Computers are everywhere except for in the productivity number,” Ozkan said, paraphrasing Nobel laureate Robert Solow, who spoke about IT improvements in the 1980s and proved again during the dot-com bubble. In both the dot-com era and the current AI hype, there was a disconnect between technological optimism and the actual economic data. In the dot-com era, the economists explained, the economy reflected the latter scenario where gains failed to show, bursting the bubble.
AI is seemingly ubiquitous, and one could reasonably assume that it’s driving economic growth. But the technology’s returns are still yet to be seen. As the authors note, TFP growth has averaged just 1.11% annually since the launch of ChatGPT in 2022. That’s below the historical average of 1.23%, according to data from the Federal Reserve Bank of San Francisco.
Still, the authors lay out two possible scenarios as to how the AI hype could impact the economy. It’s all dependent on whether or not reality eventually catches up with said hype. If the anticipated gains do materialize—if businesses become more productive thanks to AI—the economy will experience stronger output growth, which would be accompanied by declining inflation as potential output expands.
On the flip side, if those gains fail to materialize, the economy could tip into a “prolonged period of weak growth and persistently elevated inflation.”
But there are major differences between the hype cycles. For one, during the dot-com era, much of the infrastructure built—such as fiber-optic cables—remained underutilized for years. Today, there’s high demand for AI’s critical infrastructure, data centers, with vacancy rates of just 1.4%, according to commercial real estate firm CBRE. But the buildout continues, with a highly concentrated set of tech firms investing a whopping $700 billion in AI infrastructure.
But the economists caution there’s still high uncertainty hanging in the air around AI’s payoff. “We don’t really know what are going to be those productivity gains,” Faria-e-Castro said. “We don’t know when they’re going to realize—and if even they’re going to realize.”
