Technology, Automation and Inequality

Project Outcomes Statement

How are new technologies — robots, software, and now artificial intelligence — reshaping the U.S. labor market? Who benefits and who loses when machines take on tasks that workers used to perform? And what happens to the wages, employment opportunities, and economic prospects of different groups of Americans as these changes unfold? This project, supported by the National Science Foundation from 2021 to 2025, addressed these questions through a sustained program of theoretical and empirical research on the economic consequences of automation and new technologies.

The project produced a series of articles, working papers, and synthesis pieces published in leading economics journals, including the Quarterly Journal of Economics, American Economic Journal: Macroeconomics, Journal of Monetary Economics, Annual Review of Economics, and the Handbook of Labor Economics. Across these contributions, several findings stand out.

First, the project advanced our understanding of how automation affects wages and inequality. New technologies do not simply make workers more productive across the board; they often displace workers from specific tasks, with concentrated negative effects on workers whose jobs were most exposed. The project's work shows that this kind of task displacement accounts for a large share of the rise in U.S. wage inequality over the past four decades. A new finding from this period demonstrates that automation also tends to target jobs that previously paid workers more than their outside options — so-called "rent" jobs — and that this targeting represents an inefficient form of value destruction that offsets some of the productivity gains from new technologies.

Second, the project shed light on how firms differ in their adoption of automation, and what this means for the broader economy. While the aggregate share of national income going to workers has been declining for decades, the project showed that the median American firm has actually become more labor-intensive over the same period. This apparent paradox is resolved by recognizing that automation involves substantial fixed costs that only large firms can bear: large firms automate, expand, and capture market share, while smaller firms continue to operate with traditional, labor-intensive technologies. This finding has important implications for understanding rising market concentration and the changing structure of American business.

Third, the project introduced new ways to think about the economics of advanced artificial intelligence. As AI systems become more capable, fundamental questions arise about how they should be priced and deployed, what their long-run consequences are for wages and the labor share of income, and what kinds of work humans will continue to do. The project developed frameworks for analyzing these questions rigorously, including a framework for how socially responsible AI firms should price their products to balance broad access against labor-market disruption, and a framework for thinking about long-run economic dynamics in a world where compute can substitute for human labor at large scale.

Fourth, the project produced a new account of the dynamics of the college wage premium — the gap between the wages of college-educated workers and those without a college degree. The work shows that the pace of new technology creation, rather than the inherent skill bias of any particular technology, is a key driver of this premium across time, regions, and age groups. The deceleration of new technology creation since the 2000s helps explain the recent flattening of the college premium.

The project's broader impacts include the training of two doctoral students who supported the research and have now launched their academic and policy careers — one as an Assistant Professor at the University of Pittsburgh, the other as an economist at the International Monetary Fund. Both have continued to advance research on technology and labor markets. The project's results have been disseminated through publications in leading journals, the NBER working paper series, conference presentations, and an upcoming feature in the NBER Digest. The two synthesis pieces — the article in the Annual Review of Economics and the chapter in the Handbook of Labor Economics — are intended to serve as standard references for students, researchers, and policy analysts, and provide an accessible entry point into the economics of technology and work.

Taken together, the project's contributions provide rigorous, empirically grounded answers to questions that are central to economic policy in the United States and abroad — and that will only grow more pressing as artificial intelligence continues to develop.