Why artificial intelligence isn’t boosting the economy—yet
By Alex Verkhivker – Measured productivity has been declining for more than a decade in the United States and abroad. It calls to mind Solow’s paradox, a 1987 observation by the Nobel laureate economist Robert Solow, who noted that one “can see the computer age everywhere but in the productivity statistics.”
It shouldn’t be a surprise that the same thing is happening with artificial intelligence, or AI, according to MIT’s Erik Brynjolfsson, MIT PhD candidate Daniel Rock, and Chicago Booth’s Chad Syverson.
AI is a once-in-a-lifetime, general-purpose technology that promises to provide an “engine of growth,” they write. This was also true of the steam engine, electricity, and the internal combustion engine.
And yet, the researchers point out, the steam technologies that drove the US industrial revolution took nearly 50 years to show up in rising productivity statistics. And the first 25 years after the development of the electric motor and internal combustion engine were associated with a productivity slump, with growth of less than 1.5 percent a year. Then in 1915, the pace of economic expansion doubled for 10 years.
In these cases, the researchers find signs of what they call “the productivity J-curve,” a period in economic data when productivity growth is underestimated, followed by a period when it’s overestimated. This dynamic may have also applied to the computer-powered information-technology era, with 25 years of slow productivity growth followed by a decadelong acceleration, from 1995 through 2005.
Why does this happen? more>