"The Occupational Impact of Artificial Intelligence: Labor, Skills, and Polarization"

by Felten, Edward W; Raj, Manav; Seamans, Robert (2019)

Abstract

Although artificial intelligence (AI) promises to spur economic growth, there is widespread concern that it may replace human labor. We investigate the link between AI and labor by creating a new measure that we call the AI Occupational Impact (AIOI). The AIOI measure links advances in specific applications of AI, such as image recognition, translation, or the ability to play strategic games, to workplace abilities and occupations. We use this measure to study the relationship between AI and wages, employment, and labor market polarization. We provide evidence that, on average, occupations impacted by AI experience a small but positive change in wages, but no change in employment. We also provide evidence that the positive correlation with wages is driven primarily by occupations with higher software skill requirements, and that higher-income occupations have a strong positive relationship between our measure of AI impact and both employment and wages. These findings suggest that access to complementary skills and technologies may play an important role in determining the impact of AI, and that AI has the potential to exacerbate labor market polarization.

Key Passage

Technological progress in our model takes the form of an ongoing decline in the cost of computerizing routine tasks, which can be performed both by computer capital and low skill (‘non-college’) workers in the production of goods. The adoption of computers substitutes for low skill workers performing routinetasks–such as bookkeeping, clerical work, and repetitive production and monitoring activities–which are readily computerized because they follow precise, well-defined procedures. Importantly, occupations intensive in these tasks are most commonplace in the middle of the occupational skill and wage distribution. (p.5)

Keywords

Artificial Intelligence, Automation, Employment, Polarization

Themes

Automation

Links to Reference

• https://papers.ssrn.com/abstract=3368605
• https://papers.ssrn.com/sol3/Delivery.cfm?abstractid=3368605
• http://dx.doi.org/10.2139/ssrn.3368605
• https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3399180

Citation

Share

How to contribute.