����AV

Artificial intelligence seems ubiquitous. This year’s Super Bowl ads touted the latest innovations in AI. Geopolitical powers race for AI dominance. Stock prices soar as AI is portrayed as the next great industrial revolution.

Gary Smith, Fletcher Jones Professor of Economics, hears the hype and sees in the overheated stock prices another bubble in the making. He shares thoughts on how to recognize a bubble and not succumb to FOMO—the Fear of Missing Out.

You recently warned that AI is a bubble, much like the dot-com bubble in 2000. What flashing red lights do you see of an AI bubble?

The financial value of any investment—bonds, stocks, businesses, rental properties—is based on the income it generates. Speculators ignore the income (or lack thereof) and embrace the Greater Fool Theory: paying foolish prices because they expect to sell a short time later to an even bigger fool. During the dot-com bubble, fools pushed most dot-com stock prices far higher than could be justified by any plausible assumptions about income. The same is true today of most AI stocks.

What are investments you call “story stocks”?

A story stock attracts investor attention because it has a great “story.” For example, bicycles and radios were great inventions that created speculative bubbles in bicycle and radio stocks. The dominant story now is fanciful claims about ChatGPT and other large language models (LLMs). Wharton Professor Ethan Mollick has that the productivity gains from LLMs might be larger than the gains from steam power. Sundar Pichai, CEO of Alphabet and Google, LLMs are “more profound than fire” and Turing winner , “I think it’s comparable in scale with the Industrial Revolution or electricity—or maybe the wheel.”

What will keep AI from leading to “the end of work” as some have predicted?

Some AI is unquestionably useful, but the claim that ChatGPT and other LLMs will lead to the end of work anytime soon is farfetched because LLMs can’t be trusted. LLMs are very much like Nigel Richards, the greatest Scrabble player of all time, who has won several French language scrabble championships despite not understanding the meaning of the words he spells. In the same way, LLMs do not understand the meaning of the text they input and output. They consequently cannot tell the difference between truth and falsehood or the difference between correlation and causation. They lack the critical-thinking abilities needed to make trustworthy decisions.

The real danger today is not that computers are smarter than us but that we think they are smarter than us and therefore trust them to make important decisions—like evaluating job applicants, approving loans, setting insurance rates or determining prison sentences.

How could AI stocks be appropriately valued?

All stocks should be valued in the same way—by the discounted value of the companies’ current and future profits. The problem with most AI stocks is that the profits, if any, are distant and highly uncertain. During gold rushes, the most reliable profits come from selling mining equipment. Now, the most reliable profits come from selling chips and advice. Linas Beliūnas recently that:

“Fortune 500 CEOs who know nothing about AI are given a mandate by the board who knows nothing about AI to ‘do something with AI.’

So they hire consultants who know nothing about AI to give advice and write reports about AI.

The companies making the most money from AI aren’t AI firms. It’s AI consultants.”

Why do people—even experts—continue to get on bubble bandwagons?

Even highly intelligent people can be seduced by a Fear of Missing Out (FOMO). As J.P. Morgan observed, “Nothing so undermines your financial judgment as the sight of your neighbor getting rich.”

During the South Sea Bubble in the 1700s, the “story” was that the British government had granted the South Sea Company exclusive trading privileges with South America. Never mind that Spain and Portugal controlled most of South America—and England was at war with Spain.

Nonetheless, the price of South Sea stock increased so rapidly that it was said that you could buy stock as you entered Garraway’s coffeehouse and sell it for a profit on the way out. If people were eager to buy, swindlers were eager to sell. One company was formed “for carrying on an undertaking of great advantage, but nobody is to know what it is.” The shares were priced at £100 each, with a promised annual return of £100. After selling all the stock in less than five hours, the promoter left England and never returned.

Stock prices soared because people believed that prices would continue rising, at least until they could sell to the next fool in line. In the spring of 1720, Sir Isaac Newton said, “I can calculate the motions of the heavenly bodies, but not the madness of people” and sold his South Sea shares for a £7,000 profit. Later that year he bought shares again, just before the bubble burst, and lost £20,000 (about $1.5 million in today’s dollars). After James Milner, a member of the British Parliament, was bankrupted by the South Sea Bubble, he explained that, “I said, indeed, that ruin must soon come upon us but . . . it came two months sooner than I expected.”

If Newton and Milner can be blinded by a fear of missing out, so can we all.