Will artificial intelligence transform the economy? – 03/10/2023 – Paul Krugman

Will artificial intelligence transform the economy? Today, I thought I’d take a break from my usual preoccupation with ongoing crises to reflect a little on how technology could change the economic landscape in the coming years, including a topic that seems important but hasn’t received much attention: how AI could change the United States budgetary outlook.

Since the fourth quarter of last year, there has been a huge increase in buzz, both positive and negative, about AI. This buzz appears to have died down to some extent, with the use of ChatGPT, the most famous implementation of the technology, waning in recent months.

And many observers have realized that what we’ve been calling AI — or what more careful people call “generative AI” — isn’t really intelligence at all. It is, in fact, an extrapolation of pattern recognition. Or, as some people I talk to say, it’s basically a souped-up autocorrect.

But that doesn’t mean it’s not important. After all, much of what human workers, even those considered highly skilled, do for a living is also undoubtedly souped-up autocorrect. How many workers deal with creative thinking? Even among workers who do, how much time is spent recognizing patterns rather than being creative?

I don’t say this to disrespect knowledge workers, but rather to suggest that what we’re calling AI could be a big deal for the economy, even if it doesn’t lead to the creation of Hal 9000 or Skynet.

But how big? And what type of business?

Obviously, no one really knows. Some people are trying to figure out the impact from the bottom up, looking at various types of work and estimating how much of that work can be replaced or supplemented by AI. The most widely reported numbers come from Goldman Sachs, whose base scenario predicts that AI will increase the rate of productivity growth — output per hour worked — by almost 1.5 percentage points per year over a decade, totaling around 15% in this period.

Is this plausible? Actually yes. A parallel, if you’ve studied the historical relationship between technology and productivity, is the productivity boom from 1995 to 2005, which followed decades of weak productivity growth.

As a recent Brookings Institution paper points out, this boom was driven primarily by “total factor productivity” — an increase in output per unit of input, including capital.

Economists often associate total factor productivity (TFP) growth with technological progress. This is sometimes a little dubious, since the TFP is really a “measure of our ignorance”, simply the part of economic growth that we can’t explain any other way. But from 1995 to 2005, it seems pretty clear that the boom was driven by information technology.

When productivity growth slowed, productivity was about 12% higher than the previous trend. Since AI is arguably an even more profound innovation than the technologies that drove the boom from 1995 to 2005, 15% is by no means unrealistic.

But will greater productivity make us richer or simply reduce the number of jobs? Fears of technological unemployment — a term coined by John Maynard Keynes himself in 1930 — date back to at least the early 19th century.

They even inspired a very good novel, “Mechanical Piano” by Kurt Vonnegut. While technology has often eliminated some jobs, historically this has always been, as Keynes wrote, “a temporary phase of maladjustment”, with other forms of employment emerging to replace the lost jobs.

For example, the impact of Microsoft Excel—the emergence of spreadsheet programs—appears to have eliminated many accounting jobs, but these have been replaced by increased employment, even in financial analysis.

By the way, in the same essay, Keynes predicted a future in which people would work much less than in his time and in which finding rewarding ways to fill our leisure hours would become an important social concern.

The fact that this hasn’t happened in the last 90 years is reason to be skeptical of people making similar predictions now, like Jamie Dimon, who predicted the other day that AI would lead to a three-and-a-half-day work week.

However, while there is no reason to believe that what we are calling AI will lead to mass unemployment, the technology could harm people who are displaced from their jobs and have difficulty finding new ones, or are forced to accept lower wages. . Who are the potential losers?

The likely answer is that the significant impacts will be on high-level administrative jobs, many of which are currently well-paid, while manual jobs will be largely spared.

Now, while this seems about right for generative AI, there are other applications of big data that can affect manual jobs. For example, with all the buzz around ChatGPT, there has been relatively little attention to the fact that, after years of failed hype, self-driving cars are actually starting to operate.

Still, at this point, it seems likely that AI, unlike the technological progress of the last 40 years, will be a force for reducing, not increasing, income inequality.

Finally, it seems worth considering how generative AI might affect an issue that has come back to the fore: concerns about government debt.

Until recently, many economists, myself included, argued that public debt was less of a concern than many people realize because interest rates on debt were below the economy’s long-term growth rate, “r < g."

This meant that the common idea that debt would accumulate, with interest payments leading to greater debt and therefore even greater interest payments, was wrong: the debt-to-GDP ratio, the number that matters, would tend to decline instead of accumulating.

But rapidly rising interest rates have made the debt considerably more worrisome. Conventional estimates of the economy’s long-term sustainable growth rate, such as those from the Federal Reserve, tend to put it around 1.8%. And real interest rates on federal debt are now above that number.

Discussions about debt sustainability, however, are strangely disconnected from the discourse on generative AI. In fact, I’m pretty sure there are people warning of both a debt crisis and mass unemployment caused by AI, although I haven’t made the effort to find them.

But if the optimistic estimates of technology’s boost are correct, growth will be much higher than 1.8% over the next decade, and debt won’t be a big concern after all — especially since faster growth will raise revenue and reduce the budget deficit. .

All of this is, of course, highly speculative. No one really knows what the impact of AI will be. But again, it doesn’t have to be “true” artificial intelligence to be important for the economy, and the best guess is that it will likely have a big impact.