"Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction"

by Agrawal, Ajay; Gans, Joshua S; Goldfarb, Avi (2019)

Abstract

Recent advances in artificial intelligence are primarily driven by machine learning, a prediction technology. Prediction is useful because it is an input into decision-making. In order to appreciate the impact of artificial intelligence on jobs, it is important to understand the relative roles of prediction and decision tasks. We describe and provide examples of how artificial intelligence will affect labor, emphasizing differences between when automating prediction leads to automating decisions versus enhancing decision-making by humans.

Key Passage

For any given worker, a key predictor of whether artificial intelligence will substitute for their job is the degree to which the core skill they bring to the job involves prediction. Transcription jobs are being automated as the core skill of that labor is predicting which words to type upon hearing a recording. For London taxi drivers, when artificial intelligence was employed to predict the optimal route through the city’s streets, their jobs were put at risk (though other drivers’ labor became augmented). Artificial intelligence does not fit easily into existing analyses of the effect of automation on labor markets. The reasons are threefold. First, prediction is always strictly complementary to other tasks—namely decision-related tasks. Those tasks can be existing or newly possible because of better prediction. Second, better prediction improves decisions—whether taken by labor or capital—by enabling more nuanced decisions through the reduction of uncertainty. Finally, it is not yet possible to say whether the net impact on decision tasks—whether existing or new— is likely to favor labor or capital. We have found important examples of both, and there is no obvious reason for a particular bias to emerge. Thus, we caution on drawing broad inferences from the research on factory automation (for example, Acemoglu and Restrepo 2017; Autor and Salomons 2018) in forecasting the net near-term consequences of artificial intelligence for labor markets. (p.47)

Keywords

Artificial Intelligence, Labor Markets, Automation, Job Futures, Work Prediction, Machine Learning

Themes

Employment, Automation

Links to Reference

• https://www.aeaweb.org/doi/10.1257/jep.33.2.31
• http://dx.doi.org/10.1257/jep.33.2.31
• https://pubs.aeaweb.org/doi/pdfplus/10.1257/jep.33.2.31

Citation

Share

How to contribute.