The BaseResearchU.S. Census Bureau
You’re (not) hired: Artificial intelligence and early career hiring in the Quarterly Workforce Indicators
Lee C. Tucker's Census Bureau working paper compares industries by AI exposure in administrative employer-employee data. Hires of 22-24-year-olds in the most exposed fifth fell about 9% relative to others when ChatGPT was released and stayed down; their adjusted employment fell 12% over ten quarters. Tucker finds earlier trend shifts and advises some caution in attributing it all to AI.
Why it matters
The same thing, for different readers.
What it means for you
For young job-seekers, official records show fewer hires into AI-exposed industries since ChatGPT's release; the author cautions that remote work and schooling may also play a part.
For lawmakers
These data cannot separate AI from remote work or schooling; the author points to linking AI-adoption surveys with employer-employee records as the way to learn more.
For researchers
Exposure is assigned by industry, not firm; hires fall at ChatGPT's release and interest-rate sensitivity explains little of that, yet triple-difference estimates show weakening from around COVID.
The question it raises
One mechanism the author floats: uncertainty about what AI can do may lead firms to hold back hiring of young workers, an effect that would precede widespread adoption.
In their own words
Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.
From the source’s own text.
“Regression adjusted employment of early career workers in the most AI-exposed quintile of industry-state cells declined by 12% over the 10 quarters following the introduction of ChatGPT, even as employment in less exposed industries has remained stable.”
“Using detailed tabulations from matched employer-employee administrative data, I document evidence of an immediate, sizable, and persistent decrease in the level of early career (22-24 year old) hires following introduction of ChatGPT within the industry-state cells that are most exposed to AI.”
“However, triple difference estimates provide some evidence of earlier trend shifts on employment, hiring, and separations around the onset of the COVID pandemic. I discuss potential explanations, including the increase in remote work and increased educational attainment among workers in AI-exposed occupations.”
“A historical decomposition suggests that up to one quarter of relative early career employment declines through 2025q2 may be attributable to monetary policy shocks through 2023, but the analysis does not find evidence that these shocks can explain the rapid decline in hires at the most AI-exposed firms in comparison to others.”
“The results of this analysis encourage some caution in attributing all of this to AI.”
How the finding was reached
In our words, from the source. Weigh the finding by how it was made.
- What was done
- Event-study and triple-difference analysis of public Quarterly Workforce Indicators tabulations, which come from matched employer-employee administrative data covering a near-universe of private-sector jobs. Hires, separations, employment and earnings of 22-24-year-olds are compared across industry-by-state cells ranked into five bands of AI exposure (an occupation-level measure made with GPT-4 task ratings, converted to industries). A flows-based counterfactual and a local-projections test of sensitivity to monetary policy follow.
- What it found
- Hires of 22-24-year-olds in the most AI-exposed fifth of industry-state cells fell about 9% right after ChatGPT's release, relative to less exposed cells, and had not recovered; over the last four quarters they were 12.7% below the reference level. Employment of 22-24-year-olds in that fifth fell 15.2% from 2022Q4 to 2025Q2 (12.4% regression-adjusted), about 159,000 jobs, which the paper says fewer hires fully account for. The hiring rate largely recovered by early 2025 only because the employment base had shrunk. Earnings growth of early-career workers in exposed industries slowed slightly. Employment of older workers in the same industries was comparatively unaffected.
- What it does not show
- The author calls it premature to read the hiring drop as clear evidence of AI's long-run effect. Triple-difference estimates show young workers' hires, employment and separations were already weakening relative to older workers around the start of COVID, so remote work or delayed entry through schooling may play a role; a monetary-policy test suggests up to a quarter of the relative employment decline through 2025Q2 could reflect interest-rate shocks. Exposure is assigned by industry, not firm or worker, age bands are coarse, and firm-level AI adoption is not observed. Census states the paper has not had the review given to its publications.
- Status
- A working paper: not yet peer reviewed. Section 3 replicates the Stanford payroll-data result in administrative data: the author calls the unadjusted 15.2% decline highly consistent with the 16% regression-adjusted estimate in that study's 2025 version.
The record
- Questions
- What happens to work, income and the economy?
- People
- Lee C. Tucker (Wrote it, U.S. Census Bureau)
- What we read
- We read the whole source.
- Found by
- A research pass: work, income and the economy, gathered October 13, 2026, via
research-work(tier 1: the source’s own channel or an official record) - Cite this
- U.S. Census Bureau, “You’re (not) hired: Artificial intelligence and early career hiring in the Quarterly Workforce Indicators”, April 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/census-2026-04-ai-early-career-hiring-qwi