This site is a work in progress and has not been widely shared. Content may contain errors. Feedback is welcome.
This site is undergoing review. Some annotations were human-generated, some AI-generated — all are being verified.
Back to papers

Applying AI to Rebuild Middle Class Jobs

Autor

2024NBER Working Paper Series93 citations
Theoretical / conceptual
AI (General)HealthcareLegalSoftware / codingEducationAugmentation vs. substitutionTraining / upskillingJunior / entry-level
DOI: 10.3386/w32140
Summary

Autor develops a conceptual framework arguing that AI could restore middle-class employment by enabling workers with foundational training to perform expert decision-making tasks currently restricted to elite professionals, drawing on historical analysis of technological change and recent experimental evidence.

Main Finding

AI could potentially restore middle-class jobs by enabling workers with foundational expertise to perform higher-stakes decision-making tasks currently restricted to elite experts (doctors, lawyers, software engineers, professors), thereby moderating inequality and lowering costs of key services, though this outcome depends on institutional choices rather than technological determinism.

Primary Datasets

CPS

Labor outcomes

O*NET

Other

Census; CPS; O*NET

Secondary Datasets

None

Key Methods
Conceptual framework drawing on historical analysis of technological change and expertise; synthesizes recent experimental evidence from other studies
Sample Period
Historical and forward-looking
Geographic Coverage
US
Sample Size
Not applicable (conceptual/framework paper)
Level of Analysis
Occupation, Individual, Task
Occupation Classification
O*NET-SOC
Industry Classification
None
Notes
NBER WP 32140; Ely Lecture (AEA 2024) [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis. [Claude classification]: Originally drafted for NOEMA Magazine; Ely Lecture at AEA 2024. This is an essay/argument piece rather than an empirical study. Autor explicitly states 'My thesis is not a forecast but an argument about what is possible.' The paper synthesizes historical perspective on expertise (artisanal → mass → elite) with recent experimental evidence from other researchers. References to Census/CPS/O*NET are illustrative for historical context, not systematic data analysis.