As generative AI instruments turn out to be extra extensively used, a key situation is the expertise’s impression on labor demand. The place may we discover proof of that impression? On this publish, we look at whether or not early proof of AI’s impact on the labor market seems in corporations’ job postings. We mix an occupational measure of AI publicity with detailed U.S. job-posting information from Lightcast, which aggregates listings from firm profession pages, nationwide and native job boards, and job-listing aggregators. Utilizing this information, we check whether or not postings for AI-exposed occupations declined disproportionately for the reason that launch of ChatGPT in late 2022. We discover that, whereas general hiring has slowed since then, the proof from job postings offers little indication of a definite AI-driven decline in labor demand.
Measuring a Job’s Publicity to AI
To measure how uncovered completely different jobs are to AI, we use a task-level AI exposure metric developed by Anthropic that mixes detailed job descriptions from O*NET with noticed AI utilization. O*NET breaks every occupation right into a set of particular actions that staff recurrently carry out. For instance, a copywriter might edit or rewrite advertising and marketing textual content, whereas an online developer might write supporting code for web sites and net purposes. The Anthropic measure evaluates every job and assigns it an AI publicity rating primarily based on three components: whether or not the duty may theoretically be largely accomplished by AI; whether or not the duty truly seems in a pattern of AI utilization information; and whether or not AI is used to automate the task rather than augment it.
Duties obtain the next AI publicity rating if most of the noticed utilization is used to automate, moderately than increase, work. These task-level scores are then aggregated to the occupation stage utilizing data on how a lot time staff spend on every job, producing an occupation-level measure of publicity to AI on a scale from 0 to 1.
This measure must be interpreted because the potential AI publicity of an occupation primarily based on noticed utilization. A job being uncovered to AI might not translate into decreased hiring or elevated layoffs for the occupation as an entire; within the New York Fed’s Second District, considerably extra firms report retraining workers in AI-exposed occupations than reducing hiring. In observe, even when many duties inside an occupation are extremely uncovered to AI, a single task may limit the extent to which the occupation as a whole can be automated.
Utilizing this measure of an occupation’s publicity to AI, the chart beneath compares the distribution of AI publicity throughout occupations in employment (blue) and in job postings (gold). Every bar exhibits the share of staff or vacancies in occupations inside a given vary of AI publicity. Shifting proper alongside the x-axis corresponds to occupations with larger publicity, whereas the y-axis reviews the share of employment or postings in these publicity bins.
AI Publicity Stays Restricted in Each Employment and Vacancies
Notes: Bars present the share of complete employment in 2024 (blue) and vacancies from January 2026 (gold) throughout occupations grouped by ranges of AI publicity. (The darkish inexperienced is the place the 2 overlap.) The x-axis reviews bins of AI publicity, and the y-axis reviews the share of employment or vacancies inside every bin.
The chart highlights that AI publicity stays comparatively restricted. Solely a small share of employment or vacancies is concentrated in occupations with excessive AI publicity—lower than 10 p.c of staff and vacancies are in occupations with an AI publicity of no less than 0.4—and 40 p.c of staff are in jobs with zero measured AI publicity. Given this restricted publicity, will we see any impression of AI once we take a look at the change in job postings over time?
To look at whether or not AI is affecting labor demand, we conduct an occasion research that compares how job postings evolve for occupations with comparatively excessive versus low AI publicity across the launch of ChatGPT in late 2022. Right here, we outline high-exposure occupations as these with an AI publicity of no less than 0.2; the outcomes are related underneath various cutoff values.
The chart beneath plots the estimated distinction in job postings between these high-exposure occupations and less-exposed occupations for every quarter relative to the final interval previous to the discharge of ChatGPT (the distinction in 2022:Q3 is zero by building). The blue line within the chart exhibits how rather more (or much less) hiring occurred in high-exposure occupations in contrast with low-exposure occupations at every time limit, relative to the quarter earlier than ChatGPT was launched. The shaded space depicts statistical uncertainty round these estimates. This occasion research additionally accounts for persistent variations throughout occupations (since some jobs constantly have extra postings than others) and economy-wide modifications in hiring over time, permitting us to concentrate on variations in hiring by AI publicity.
Declines in Vacancies for AI-Uncovered Occupations Started Earlier than the Launch of ChatGPT in Late 2022
Notes: Occupations are categorized as “excessive publicity” if they’ve a job-level publicity of no less than 0.2. Occupation weights are derived from the quantity of vacancies for that occupation in 2019. The vertical pink line signifies the quarter of ChatGPT’s first public launch. Shaded areas point out 95 p.c confidence intervals.
If AI had had a major causal impact on employment, we’d count on the employment distinction between uncovered and fewer uncovered occupations to behave within the following two methods. First, previous to ChatGPT’s launch, hiring developments in high- and low-exposure occupations would transfer equally. This might counsel that, within the absence of AI, the 2 teams would have continued evolving equally. Within the chart, this is able to correspond to estimates being statistically indistinguishable from zero in all quarters previous to 2022:Q3. Second, a sustained divergence between high- and low-exposure occupations ought to emerge sooner or later after ChatGPT’s launch. A spot that opens up—and particularly one which grows over time—could be in step with AI affecting labor demand.
Whereas the chart exhibits a relative decline in postings for occupations with larger AI publicity, the occasion research signifies that this pattern predates the discharge of ChatGPT. The divergence between high- and low-exposure occupations started earlier than 2022 and doesn’t present a transparent extra break in trajectory after 2022. Moreover, the hole in labor demand between high- and low-exposure jobs stabilizes after 2023, at odds with AI step by step displacing uncovered occupations. This makes it troublesome to interpret the relative decline in hiring in AI-exposed occupations as a direct consequence of AI adoption.
Is AI Lowering Demand for Entry-Stage Jobs?
A lot of the early dialogue about AI’s labor-market results has centered on youthful and entry-level staff. Analysis on the employment impression of AI has discovered a bigger decline within the quantity of youthful staff in occupations with excessive AI publicity after the discharge of ChatGPT. On the identical time, associated work utilizing job postings finds that demand for junior and senior roles in these occupations declined at roughly the identical time and by related magnitudes starting in 2022.
We conduct one other occasion research to measure the distinction in postings between junior and senior roles inside occupations with excessive AI publicity, relative to late 2022 (proven within the chart beneath). Values above zero point out that postings for junior roles elevated relative to these for senior roles inside the identical high-AI-exposure occupation, whereas values beneath zero point out the other. For instance, if the road remained constantly above the horizontal axis after 2022, it might counsel that labor demand for junior positions in high-AI-exposure occupations had grown relative to hiring for senior roles in those self same occupations.
No Clear Divergence in Labor Demand Between Junior and Senior Positions in Occupations with Excessive AI Publicity
Notes: Occupations are categorized as “excessive publicity” if they’ve a job-level publicity of no less than 0.2. Job postings are categorized as both “junior” or “senior” stage by Lightcast primarily based on data within the posting. Occupation weights are derived from the quantity of vacancies for that occupation in 2019. The vertical pink line signifies the quarter of ChatGPT’s first public launch. Shaded areas point out 95 p.c confidence intervals.
If AI had been disproportionately lowering demand for entry-level work, we’d count on the road to maneuver downward after 2022, indicating a relative decline in postings for junior roles. As an alternative, the road fluctuates, with no clear upward or downward pattern. This implies that labor demand for junior and senior roles inside extremely uncovered occupations is shifting broadly in parallel, and that the slowdown in postings is just not concentrated particularly in entry-level extremely uncovered jobs.
Conclusion
General hiring has slowed since 2022, and unemployment has increased among young workers and recent college graduates. The proof from job postings means that whereas AI could also be contributing to current labor market developments, it’s not the principle driver of the slowdown in hiring. In step with this interpretation, the New York Fed’s enterprise surveys point out that, thus far, corporations intend to include AI primarily via retraining, with restricted results on hiring. Whereas job postings present a relative decline in vacancies in occupations with higher publicity to AI, that divergence started earlier than the discharge of ChatGPT in late 2022. Furthermore, we don’t observe a divergence in labor demand between junior and senior positions inside extremely uncovered occupations. These patterns make it troublesome to attribute the current slowdown in entry-level hiring to AI alone.
Richard Audoly is a analysis economist within the Federal Reserve Financial institution of New York’s Analysis and Statistics Group.
Miles Guerin is a analysis analyst within the Federal Reserve Financial institution of New York’s Analysis and Statistics Group.
Giorgio Topa is an financial analysis advisor within the Federal Reserve Financial institution of New York’s Analysis and Statistics Group.
Learn how to cite this publish:
Richard Audoly, Miles Guerin, and Giorgio Topa, “Do Job Postings Show Early Labor‑Market Effects of AI?,” Federal Reserve Financial institution of New York Liberty Road Economics, Might 14, 2026, https://doi.org/10.59576/lse.20260514
BibTeX: View |
Disclaimer
The views expressed on this publish are these of the creator(s) and don’t essentially mirror the place of the Federal Reserve Financial institution of New York or the Federal Reserve System. Any errors or omissions are the accountability of the creator(s).
