AI and the Entry-Level Labor Market
Did early-career employment change differently after widespread generative-AI adoption in occupations with greater measured AI exposure?
- Status
- Active development · no results yet
- Role
- Coauthor with Shane Vardanyan
- Authors
- Ian Helfrich, Shane Vardanyan
- Method
- continuous-exposure event study · triple difference · Current Population Survey microdata
- What this record does not establish
- No empirical results exist yet; the public record describes a design and planned analysis only.
A public manuscript is not yet available.
Abstract
Summary
Develops an empirical design for asking whether early-career employment changed differently in occupations with greater generative-AI exposure. The planned analysis combines continuous-exposure event-study and triple-difference specifications using Current Population Survey microdata. The paper is in active development and claims no empirical result yet.
Methods and notes
This project asks whether the diffusion of generative AI has changed employment outcomes differently for workers entering the labor market than for more experienced workers, and whether those changes vary systematically with occupational exposure to AI.
The design uses Current Population Survey microdata and two complementary approaches: a continuous-exposure event study and a triple-difference specification comparing higher- and lower-exposure occupations across experience groups before and after widespread generative-AI adoption.
The paper is in active development. A public draft, full variable construction, robustness analysis, and replication materials will be posted when the estimates are ready for circulation. No results are claimed on this page before that stage.