The AI revolution may be upon us. But nobody knows for sure what it will mean for measures of overall economic well-being like GDP, unemployment, incomes and asset prices.
The big picture: Will super-intelligent computing unleash stunning growth in output — or will it be more like some previous big advances that created "ghost GDP" that didn't show up much in the economic statistics?
- Will it push millions of workers — especially white-collar professionals with high incomes — onto the unemployment rolls? Or enable them to be far more productive and thus earn more?
- Will it push more of the national income toward owners of capital and away from human laborers? If so, what does that mean for the broader perpetual motion machine that is the economy, where one person's spending becomes another person's income?
- Does it justify stratospheric prices for stocks, given the potential savings?
- Or should it terrify investors who could see consumer demand dry up and now-lucrative business models implode?
Zoom out: The sheer range of possibilities is dizzying.
- These discussions have been roiling beneath the surface among technologists and economists. Now, they are getting more attention — and not as a distant dystopia, but as near-term risks.
State of play: On Sunday, Citrini Research and Alap Shah published a long, speculative essay that amounts to a doom-and-gloom scenario of where the economy and markets could find themselves in just two years.
- In their scenario, companies rush to replace human employees with AI tools, only to soon find that demand for their products evaporates just as quickly.
- This causes a 38% sell-off in stocks and problems in credit markets like mortgages, as displaced workers can't pay their loans.
- "Hopefully, reading this leaves you more prepared for potential left tail risks as AI makes the economy increasingly weird," the authors wrote.
Yes, but: There are plenty of questionable assumptions in their scenario. One is that companies will be able to cut back on human employment as rapidly as assumed. Another is that a rise in unemployment wouldn't be matched by aggressive action out of Washington.
- Still, it is a sign of the times that this 7,000-word essay about a hypothetical future has gone mega-viral, with the social media post that launched it getting 19 million views in a day.
Flashback: History is filled with examples of innovations that had enormous consequences for daily life and created structural economic change — but without showing up much in the data source for GDP and other statistics.
- For example, broadcast television remade families' lifestyles and the way products were marketed. But it was a free product that only appeared in statistical tables via TV purchases and spending on advertising.
- Workplace software created disruptions to clerical jobs that dwarfed spending on Lotus 1-2-3 or Microsoft Word.
Between the lines: Even if you correctly project the evolution of a technology, it doesn't follow that you can predict the economic ripple effects.
- Unfortunately, people setting economic policy don't have the luxury of sitting and waiting for the future to reveal itself. They have to set policy based on their best guess of where the puck is going.
- In some cases, that means AI-affected economic policy choices that are set to be made this year.
- The Federal Reserve may need to consider an argument put forward by President Trump's nominee to lead it, Kevin Warsh, who contends that the AI-driven productivity surge gives the Fed room to cut interest rates further without the risk of inflation.
Reality check: "Thinking through all of the different risks from a new shock like AI is helpful, necessary, and interesting," Ernie Tedeschi, the chief economist at Stripe, said in a post on X.
- "Still, my default assumption to the question, 'Will this shock be radically different from all the prior ones' is that the likeliest answer is 'No, not fundamentally.'"
