The Future of Employment
The paper by Carl Benedikt Frey and Michael A. Osborne investigates how vulnerable US occupations are to computerisation. It begins with a historical literature review of technological change, skill‑biased technical change, and previous automation forecasts. The authors then introduce a novel methodology that combines ONET occupational attributes with a Gaussian process classifier to predict the probability that each of 702 occupations can be automated. The results show that roughly 47 % of US employment is at high risk, with a clear polarization: routine, low‑skill jobs face the highest probabilities, while occupations requiring high levels of creativity, social interaction, or complex perception remain relatively safe. The paper discusses the implications for wages, education, and labor‑market dynamics, emphasizing the need for policy responses such as upskilling, education reform, and safety‑net redesign. An extensive reference list documents the interdisciplinary sources used, and an appendix provides a ranked table of occupations with their estimated computerisation probabilities, including binary labels for the training set. The work serves as a benchmark for future research on the economic impact of AI and robotics.
Topics
Historical Context: Technological Change and Skill‑Biased Growth
A review of economic history—from guilds and early industrial revolutions to modern AI advances—that frames expectations about automation and skill‑biased technical change.
Methodological Innovation: Gaussian Process Classification Using O*NET Occupational Attributes
A novel probabilistic model that combines O*NET task, skill, and technology variables with a Gaussian process classifier to estimate the probability of computerisation for each of 702 U.S. occupations.
Core Findings: Distribution of Automation Risk and the 47 % High‑Risk Estimate
The headline result that roughly 47 % of U.S. jobs face a high probability of automation, revealing a U‑shaped risk curve across wage and education levels and a clear polarization between routine low‑skill jobs and occupations requiring creativity, social interaction, or complex perception.
Economic Implications: Effects on Wages, Education, and Labor‑Market Polarization
An analysis of how automation risk correlates with lower wages and lower educational requirements, driving a split between high‑skill, high‑pay jobs and the erosion of low‑skill, low‑pay employment, and discussing potential wage pressure and skill gaps.
Policy Recommendations and Future Research Directions
Proposes policy responses such as upskilling, education reform, and safety‑net redesign, and outlines avenues for extending the model to other economies, sectors, and dynamic labor‑market analyses.