The rise of artificial intelligence (AI) undoubtedly impacts labor markets, leading to frequent discussions of whether machines will replace the human labor force or which skills individuals should invest in to be attractive for firms. This influence on labor markets may also have an effect on regions; including, how or if industries that demand AI skills agglomerate and what kind of individuals these firms attract. Hanson (2021) therefore seeks to understand regional changes in the USA due to demand for AI-related employment for the years 2000 to 2018.
The analysis breaks down into three steps. Their analysis first identifies AI-related employment based on Census-defined job titles that contain one or more of the following terms: computer, data, or software; and design/designer, engineer, research/researcher or science/scientist. The amount of new job titles within an occupation for the targeted time period are then used to measure employment growth in AI. Second, regional specialization of AI-related occupations is examined and ultimately finds two patterns: (i) regional hubs for tech jobs during the 1980/1990s gives way for being the largest “commuting zones” for AI-related jobs in 2000-2018 and (ii) foreign-born men are the primary reason these commuting zones see an increased specialization in AI-related jobs (leading to a 54.6% increase in hours worked). Additionally, the commuting zones that have a higher concentration of foreign-born men in AI-related jobs tend to be dominated by private firms; whereas, commuting zones with a higher concentration of USA-native born men in AI-related jobs tend to be focused on government-funded military research. Lastly, the analysis seeks to understand the factors that cause the regional growth for AI-related jobs.
The main findings provide that foreign-born workers neither crowd out or crowd in native-born workers, as seen in the strong, positive correlation between men and employment growth in AI-related occupation, which can be completely attributed to the increased amount of foreign-born men in AI-related jobs. This result is similar for women. Further, many of the foreign-born workers in the US arrive from India and China, and choose to settle in locations where previous generations from their native countries chose to settle. Additionally, the findings specify that the US regions most equipped to attract talent for AI-related jobs are those that invest in their comparative advantage in AI. Lastly, the findings of this paper suggest that AI-related job growth is constrained by access to high-skilled workers, but that this constraint is alleviated by the immigration of college-educated, foreign-born workers.
Hanson, G. (2021). Immigration and regional specialization in ai. In Robots and AI (pp. 180-231). Routledge.