Michael Spence has great hopes for AI

Nobel Laureate Michael Spence sees  AI as offsetting two big trends that are working against global growth.

One negative trend is that efficiency or cost has ceased to be the primary consideration in determining the source of supply:

The first is shocks, including war, pandemic, climate change, geopolitical tensions, resurgent nationalism, and growing focus on national security in the conduct of international economic policy. These increasingly severe and frequent disruptions are shifting global supply networks toward greater diversification and resilience. But that is an expensive pressure and a contributor to inflationary pressures.

Another is the fall in productivity growth:

Productivity deserves special attention. US productivity growth averaged 1.68 percent from 1998 to 2007, a period during which many Americans got internet access and, later, mobile phones. Productivity growth then slowed to 0.38 percent from 2010 to 2019.....In Europe, lagging growth and productivity are attributable in part to less rapid and effective adoption and deployment of digital technologies, and to underdeveloped tech sectors relative to the US and China.

These two forces are impacting economies in a number of ways:

The combined effect of these two sets of forces is a relatively rapid shift from demand-constrained to supply-constrained growth. Growth is subdued. Inflation endures. Real interest rates remain elevated. Many economists, including me, believe that the structural conditions I’ve described mean borrowing costs are likely to remain elevated, and certainly higher than during the decade following the global financial crisis.

Spence sees AI as counteracting these two negative forces and leading to a surge in productivity although this will take a long time to happen- he sees the impact no earlier than the towards the end of the present decade:

Of course it will take time. Roy Amara’s law applies here as in past episodes of technological transformation: we tend to overestimate the short-run impacts and underestimate the longer-term ones. My best guess (and it is just a guess, based on current patterns of investment) is that we may start to see meaningful impacts in labor productivity by the end of this decade.

Spence's views are worth noting because economists, in general, remain sceptical as to whether AI will cause productivity to accelerate- they see it as maintaining historical rates of productivity at best.