Projections about Artificial Intelligence and Employment

Technology has a fortunate history of improving welfare. Despite the fears of generations past, new waves of technology have historically displaced some workers but on average maintained or even grown the employment rate, complementing rather than replacing human labor. There is, however, an active debate about whether we should expect this trend to continue, as an increasing number of economic and computer science scholars believe that future technological development, in the form of artificial intelligence, is likely to break the fortunate trend.

In this paper I briefly describe artificial intelligence and consider both sides of the debate. Then, taking the pessimist’s view of job displacement, I try to lay the framework for quantitatively predicting the impact of increasing automation on the labor market. I borrow from prediction science literature, focusing on describing the scope, meaning, and timeline of the technology’s development. In this context, this describes the kinds of skills of which artificial intelligences are likely to be capable, how these skills appear in jobs, and when such technologies are likely to exist and become adopted in the workplace. I then break each section into the five factors the McKinsey Global Institute deems necessary conditions for the adoption of artificial intelligences in the place of humans, and discuss them in turn. I conclude by applying the framework to the example occupation of wait-staff, making predictions while voicing reservations about the predictions and areas for further investigation.


Full paper:

Projections about AI and Employment_ Working Paper