Erik Brynjolfsson
Stanford economist directing the Digital Economy Lab; author of the 34-to-1 AI displacement ratio finding in April 2026.
Last refreshed: 16 April 2026 · Appears in 1 active topic
What did Brynjolfsson find when he applied JOLTS data to AI job losses?
Timeline for Erik Brynjolfsson
Led research calculating AI's hidden displacement via hiring-rate suppression
AI: Jobs, Power & Money: Stanford: AI costs 34 hires per layoff- What is Erik Brynjolfsson's finding on AI job losses?
- Brynjolfsson's Stanford lab calculated that AI suppresses roughly 950,000 to 1 million annual hires in the US, a 34-to-1 ratio against declared AI layoffs, using February 2026 JOLTS data.Source: Stanford Digital Economy Lab
- Is The Second Machine Age prediction coming true?
- Brynjolfsson and McAfee argued digital technology would reshape labour markets faster than workers could adapt. April 2026 JOLTS data shows workers aged 22–25 in AI-exposed roles are 16% below 2022 employment levels, suggesting the prediction is now measurable.Source: Stanford Digital Economy Lab
- Why does JOLTS show more AI job losses than layoff announcements?
- JOLTS measures hires, not just cuts. Brynjolfsson applied the gap between the 3.1% February 2026 hiring rate and the 2023 baseline to the full nonfarm workforce, finding AI suppresses 34 positions for every publicly declared layoff.Source: Stanford Digital Economy Lab / ProCap Insights
Background
Erik Brynjolfsson is director of the Stanford Digital Economy Lab and one of the most influential economists studying AI's effects on labour markets. In April 2026 his team applied the JOLTS hiring rate — 3.1% in February 2026, the lowest since April 2020 — to the US nonfarm workforce of 158.6 million and concluded that AI is suppressing roughly 950,000 to 1 million annual hires against the 2023 pace. Against the Challenger, Gray & Christmas tally of 27,645 declared AI layoffs through March, that produces a ratio of roughly 34 suppressed hires per declared layoff.
Brynjolfsson co-authored The Second Machine Age (2014) with Andrew McAfee, arguing that digital technologies would reshape labour markets faster than workers could adapt. He has tracked AI adoption metrics since the mid-2010s and served on government advisory panels on automation. His core argument — that productivity gains from AI accrue faster to capital than to labour — is now supported by age-specific employment data showing workers aged 22 to 25 in AI-exposed occupations 16% below late-2022 levels, while colleagues over 30 are up 6 to 12%.
The April 2026 finding landed the same week three federal surveys produced conflicting adoption rates of 18%, 41% and 78% for the same economy, and the Bureau of Labor Statistics skipped its scheduled GenAI workplace publication. Brynjolfsson's JOLTS analysis is now the closest approximation to a federal displacement baseline that US policymakers have, making him a central figure in the data fight the Hawley-Warner senate Coalition is pressing and the Leading the Future super PAC is spending million to forestall.