Award Recipients for Volume 14
Exploring Machine Learning’s Contributions to Economic Productivity and Innovation
What role does computational power play in economic productivity and innovation? How will machine learning and AI change this? Building off previous work quantifying historical computational power levels, the paper explores the relationship of metrics for computing power with US GDP and US internet sector GDP from 1960 to today. The paper develops forecast scenarios incorporating machine learning development using internet data production volumes and forecasted growth of computational power. The goal of the research is not to build a full model of productivity that incorporates computational power, but to begin to get a sense of potential role and impacts of artificial intelligence on the economy. The research has three main findings. First, the paper finds a modest but statistically significant relationship between computational power and economic productivity, linked to approximately 0.3–0.7 percent of GDP per capita and to approximately 2–3 percent of internet sector GDP per capita. Second, and as expected, the relationship is stronger for internet sector GDP per capita, which is linked more closely to AI. Third, and as expected, when the paper narrows its window of analysis to more recent windows of analysis in the regressions and in robustness tests, it sees a strengthening of the relationship between computational power and GDP per capita.
Our article, “Exploring Machine Learning’s Contributions to Economic Productivity and Innovation,” represents an attempt to look at the socioeconomic impacts of technology through a different kind of research lens. We analyze ‘non-traditional’ metrics to better understand if and how we can link the power of computing technology to economic growth over time. The results are fascinating.
The article is exploratory in nature and serves as a starting point for updating how we think about research on technology. And it reveals several key insights. First, the article finds a clear, statistically significant linkage between computational power and economic productivity. Historically, computational power has been linked to about 0.3-0.7% of GDP since the 1960s. Second, the article finds the linkage is stronger (as expected) for the productivity of digital industries. We find computational power has been linked to approximately 2—3% of the GDP contributions of the “internet sector.” Third, these relationships have clearly grown stronger over time, meaning that computational power’s importance to the economy has become more important in recent years. Using forecasts for continued computational power growth through machine learning and artificial intelligence technologies, the economic importance of computing power to the U.S. economy will grow dramatically in the coming decades.
We also argue the article is important for other reasons beyond the direct research conclusions. The rapid growth and development of digital technologies, and the industries they buttress, have left many research fields behind. There is now a need for new ways of thinking about technology and how to research its impacts on society. We argue we need to consider different approaches and different metrics beyond those used in the 20th century. By combining fields – for example by drawing on engineering metrics as economic variables – we can potentially unlock new insights. Research should follow the technology sector’s example and strive to innovate as well.
We hope this article helps chip away at the lack of research on the socioeconomic impacts of cutting-edge technology and that the ideas explored within it might help others develop new approaches to research as well.
—Christopher Alex Hooton and Davin Kaing