We are in the midst of an economic transformation made possible by the development and rapid improvement of artificial intelligence. While we have not yet seen much of what AI promises to deliver, there is an enormous amount of investment being made on the expectation that AI will generate more than enough profits to justify those investments. At the same time, many people are worried that those profits will come at their expense. They have reason to be worried.

In 2019, Daron Acemoglu and Pascual Restrepo published an article entitled “Automation and New Tasks: How Technology Displaces and Reinstates Labor” in the Journal of Economic Perspectives. Their analysis suggests that since 1987 there has been a marked increase in the displacement of labor by automation relative to the increase in employment due to automation. Consequently, the way investments in automation of tasks has been distributed has contributed to a declining share of the nation’s income accruing to labor.

This was not inevitable. Firms found it more profitable to invest in automation that displaces workers than in machines and computer applications that enhance productivity of existing workers and create demand for new workers — in part because they can save the costs of employer-paid fringe benefits and taxes employers must pay into the Social Security and Medicare Trust Funds.

When workers are displaced, they typically find new employment that pays lower wages and fringe benefits. Not only is this a loss for those workers, but it is also a loss of revenue for governments from wage taxes that support Social Security and Medicare.

Advertise with us

These are real costs of automation, but they are not counted by the investing firm as part of their investment. A tax on investments that displace workers would correct the bias and direct automation away for job displacement and towards job creation.

JACK OCHS

Point Breeze

The writer is professor of economics, emeritus, at the University of Pittsburgh.