External signal·Liberty Street Economics (Federal Reserve Bank of New York)·May 14, 2026·8 min read
Do Job Postings Show Early Labor-Market Effects of AI?
“it is not the main driver of the slowdown in hiring”
Summary
NY Fed economists (Audoly, Guerin, Topa) combine an occupational AI-exposure measure with Lightcast US job-posting data to test whether postings for AI-exposed occupations fell disproportionately after ChatGPT. They find a relative decline in vacancies for highly exposed occupations — but the divergence began before late 2022, and there is no gap between junior and senior demand within exposed occupations.
Predictions for the future of work
Concludes AI may be contributing at the margin but is not the main driver of the hiring slowdown so far, and that firms intend to adopt AI mainly through retraining, with limited near-term hiring effects. Complicates the junior-versus-senior displacement story by showing it is not yet visible in postings.
Originally published by Liberty Street Economics (Federal Reserve Bank of New York) · May 14, 2026
Read the original at Liberty Street Economics (Federal Reserve Bank of New York)