Working Paper: NBER ID: w24690
Authors: Ajay K. Agrawal; Joshua S. Gans; Avi Goldfarb
Abstract: Recent progress in artificial intelligence (AI) – a general purpose technology affecting many industries - has been focused on advances in machine learning, which we recast as a quality-adjusted drop in the price of prediction. How will this sharp drop in price impact society? Policy will influence the impact on two key dimensions: diffusion and consequences. First, in addition to subsidies and IP policy that will influence the diffusion of AI in ways similar to their effect on other technologies, three policy categories - privacy, trade, and liability - may be uniquely salient in their influence on the diffusion patterns of AI. Second, labor and antitrust policies will influence the consequences of AI in terms of employment, inequality, and competition.
Keywords: Artificial Intelligence; Economic Policy; Inequality; Labor Market
JEL Codes: L86; O3
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Government policies related to privacy, trade, and liability (K24) | Diffusion of AI technologies (O33) |
| Diffusion of AI technologies (O33) | Economic impacts (employment patterns and income inequality) (F66) |
| Quality-adjusted drop in the price of prediction due to AI (E30) | Enhanced productivity (O49) |
| Quality-adjusted drop in the price of prediction due to AI (E30) | Changes in labor dynamics (J29) |
| Labor and antitrust policies (L49) | Consequences of AI (employment patterns and income inequality) (F66) |
| Diffusion of AI (O33) | Increased inequality (F61) |
| Government policies (H59) | Economic impacts (employment patterns and income inequality) (F66) |