AI Labor Market Tracker: August 2026
of workers have at least one AI skill, as of July 2026
Fewer layoff announcements in July 2026 at the most AI-exposed firms than the least-exposed, since Oct 2022
Decline in the number of new AI-adopting firms per month, down from its April 2026 peak
of how work is changing happens inside jobs, instead of a change in the job mix
job postings needed to make one hire
Artificial intelligence is beginning to reshape labor markets, but its effects cannot be captured by a single measure. In this second edition of the AI Labor Market Tracker, we update our recurring set of indicators tracking how AI is changing labor supply, employer demand, employment and wages, the activities performed within jobs, and the process through which workers and employers find one another.
This month, the clearest new signals are a slowdown in the pace of new firm AI adoption, continued weakness in junior high-exposure roles, and evidence that most changes in work content are occurring within occupations. At the same time, the most AI-exposed firms continue to see fewer layoffs than the least-exposed firms, underscoring that the labor-market effects of AI do not all point in the same direction.
A central challenge in studying AI's labor market effects is distinguishing anticipation from adoption. Firms may adjust hiring plans before deploying AI tools, while workers may change skill investment decisions in response to expectations about future demand. Conversely, measurable labor market effects may emerge only after organizations formally adopt AI technologies. We therefore continue to track both exposure-based measures that capture anticipated effects and firm-level adoption measures that capture realized deployment.
The purpose of the tracker is not to provide a definitive verdict on whether AI will ultimately increase or decrease employment, but to build a consistent empirical record of how the labor market is evolving as AI diffuses. Each month, we update the same core indicators while adding new analyses where the data can shed light on emerging questions. Some measures will move considerably, others may barely change, and those differences are themselves part of the picture.
- SupplyHow AI is impacting labor supplyThe share of workers reporting AI skills keeps climbing, even as Computer Science enrollment has passed its recent peak.
- DemandHow AI is impacting labor demandHiring demand has weakened in highly AI-exposed occupations, particularly at junior levels.
- Equilibrium: Employment and wagesHow AI is affecting employment and wagesEmployment is growing more slowly in exposed occupations, while AI-adopting firms continue to expand.
- Work activitiesHow AI is changing work activitiesMost of the work inside occupations is changing within jobs rather than across them, while the share of job postings with high AI exposure is declining.
- MatchingHow AI is changing the hiring processJob postings are producing fewer hires, while candidates report more recruiter ghosting and poor communication.
Computer Science enrollment peaked in 2022, even as a growing share of workers report holding AI skills on their profiles. This section tracks how students and workers are upskilling for AI.
Loading Figure 1.1 — Bachelor's-degree field shares at selected U.S. schools, as a share of all degree records each year. Source: Revelio Labs workforce data from professional online profiles.Figure 1.1 — Bachelor's-degree field shares at selected U.S. schools, as a share of all degree records each year. Source: Revelio Labs workforce data from professional online profiles.
AI skills are identified using reported skills from individual-level profile data. A worker is classified as AI-skilled if they list at least one of approximately 80 AI- and machine-learning-related skills on their profile, spanning foundational concepts (machine learning, deep learning, neural networks), generative AI tools (large language models, ChatGPT, prompt engineering), frameworks (PyTorch, TensorFlow, Hugging Face, LangChain), and infrastructure (MLOps, AWS SageMaker, MLflow). The share shown reflects the fraction of US job positions held by workers with at least one such reported skill, smoothed with a 3-month trailing average.
Loading Figure 1.2 — Share of US job positions held by workers with at least one reported AI skill. Source: Revelio Labs workforce data from professional online profiles.Figure 1.2 — Share of US job positions held by workers with at least one reported AI skill. Source: Revelio Labs workforce data from professional online profiles.
One of the most visible and established patterns from AI on the labor market is an anticipatory one: demand for AI-exposed occupations has weakened disproportionately relative to less-exposed occupations, whether or not employers have deployed any AI tools. AI exposure measures a potential rather than realized automation: the share of activities in a job that AI can credibly do. Job postings in the most exposed occupations have fallen relative to less exposed ones since late 2022 — and the effect is heavily concentrated at junior seniority levels. The results below follow the event-study design of Brynjolfsson, Chandar and Chen (2025). This section considers employer demand from job postings.
Loading Figure 2.1 — Event-study estimate of the change in posting volumes for the most AI-exposed occupations compared to the least exposed (highest versus lowest quintile exposure occupations), relative to October 2022. Two-way fixed effects regression (occupation and month), standard errors clustered by occupation. Using Revelio Labs AI exposure score. See appendix for robustness with other AI scores. Source: Revelio Labs job postings data.Figure 2.1 — Event-study estimate of the change in posting volumes for the most AI-exposed occupations compared to the least exposed (highest versus lowest quintile exposure occupations), relative to October 2022. Two-way fixed effects regression (occupation and month), standard errors clustered by occupation. Using Revelio Labs AI exposure score. See appendix for robustness with other AI scores. Source: Revelio Labs job postings data.
Loading Figure 2.2 — Event-study estimate of the change in posting volumes for the most AI-exposed occupations compared to the least exposed (highest versus lowest quintile exposure occupations), relative to October 2022; by seniority. Two-way fixed effects regression (occupation and month), standard errors clustered by occupation. Junior: Revelio seniority levels 2–3. Senior: 5–7. Using Revelio Labs AI exposure score. See appendix for robustness with other AI scores. Source: Revelio Labs job postings data.Figure 2.2 — Event-study estimate of the change in posting volumes for the most AI-exposed occupations compared to the least exposed (highest versus lowest quintile exposure occupations), relative to October 2022; by seniority. Two-way fixed effects regression (occupation and month), standard errors clustered by occupation. Junior: Revelio seniority levels 2–3. Senior: 5–7. Using Revelio Labs AI exposure score. See appendix for robustness with other AI scores. Source: Revelio Labs job postings data.
Exposure tells one story and adoption tells another. Employment is growing more slowly in occupations containing the most AI-exposed work — an anticipatory pattern. At the same time, firms that have actually adopted AI continue to expand employment overall, although the gains appear more concentrated in senior roles — a realized-adoption pattern. We measure adoption using language from job postings, following Hosseini Maasoum and Lichtinger (2025).
These patterns are not contradictory: employers across the economy can adjust hiring in anticipation of what AI may soon be able to do, while the specific firms that successfully deploy the technology grow and reorganize their workforces.
We compare employment trends in occupations with different levels of AI exposure. This approach builds on the event-study approach of Brynjolfsson, Chandar, and Chen's (2025) Canaries in the Coal Mine. Exposure measures how much of the work in an occupation current AI systems could plausibly perform; it does not indicate whether individual employers or workers have adopted AI. The results therefore capture potential anticipatory effects: employers may adjust hiring because of what they expect AI to do, before they have fully deployed the technology. This section also looks at general employment in AI-related roles, spanning all roles that touch AI; ranging from AI engineers to data center construction workers.
Loading Figure 3.1 — Headcount in AI-related roles compared to all other roles indexed to November 2022. Source: Revelio Labs workforce data from professional online profiles.Figure 3.1 — Headcount in AI-related roles compared to all other roles indexed to November 2022. Source: Revelio Labs workforce data from professional online profiles.
Loading Figure 3.2 — Event-study estimate of employment in the most AI-exposed occupations compared to the least exposed, relative to October 2022. Top versus bottom quintile AI exposure, using Revelio Labs AI exposure score. Two-way fixed effects regression (occupation and month), standard errors clustered by occupation. See appendix for robustness with other scores. Source: Revelio Labs workforce data from professional online profiles.Figure 3.2 — Event-study estimate of employment in the most AI-exposed occupations compared to the least exposed, relative to October 2022. Top versus bottom quintile AI exposure, using Revelio Labs AI exposure score. Two-way fixed effects regression (occupation and month), standard errors clustered by occupation. See appendix for robustness with other scores. Source: Revelio Labs workforce data from professional online profiles.
Loading Figure 3.3 — Event-study estimate of employment in the most AI-exposed occupations compared to the least exposed, relative to October 2022; split by age, isolating early-career workers aged 22–25. Top versus bottom quintile AI exposure, using Revelio Labs AI exposure score. Two-way fixed effects regression (occupation and month), standard errors clustered by occupation. See appendix for robustness with other scores. Source: Revelio Labs workforce data from professional online profiles.Figure 3.3 — Event-study estimate of employment in the most AI-exposed occupations compared to the least exposed, relative to October 2022; split by age, isolating early-career workers aged 22–25. Top versus bottom quintile AI exposure, using Revelio Labs AI exposure score. Two-way fixed effects regression (occupation and month), standard errors clustered by occupation. See appendix for robustness with other scores. Source: Revelio Labs workforce data from professional online profiles.
Loading Figure 3.4 — Difference-in-differences estimate of WARN layoff notices for the most versus least AI-exposed firms. Source: Revelio Labs layoff notice data from WARN.Figure 3.4 — Difference-in-differences estimate of WARN layoff notices for the most versus least AI-exposed firms. Source: Revelio Labs layoff notice data from WARN.
Quantities — headcount growth at adopting firms
Following Hosseini Maasoum and Lichtinger (2025), we identify AI-adopting firms from job postings. A firm is considered AI adopting, when it posts jobs for AI integrator roles. Note that this approach in identifying adopters is different from Kharazian, Simon and Stevens (2026), which identified adoption from spending data. Results are broadly similar.
Over the period shown, adopting firms grow headcount 26% more than non-adopting firms. Adopting firms were growing faster pre-adoption. The growth is uneven by seniority. Senior headcount grows by 32%, compared with only 6% for junior roles. While lower compared to senior level employment growth, junior growth is still higher at adopting firms compared to non-AI adopting firms.
Loading Figure 3.5 — New AI-adopting firms each month, and the cumulative share of adoption across a rolling panel of eligible U.S. hiring firms. Firms enter the panel with at least 20 new U.S. positions in the preceding 48 months. Source: Revelio Labs job postings and workforce data.Figure 3.5 — New AI-adopting firms each month, and the cumulative share of adoption across a rolling panel of eligible U.S. hiring firms. Firms enter the panel with at least 20 new U.S. positions in the preceding 48 months. Source: Revelio Labs job postings and workforce data.
Loading Figure 3.6 — Change in the headcount gap between AI integrator adopters and non-adopters relative to October 2022, with month fixed effects. Source: Revelio Labs workforce data and classified job postings.Figure 3.6 — Change in the headcount gap between AI integrator adopters and non-adopters relative to October 2022, with month fixed effects. Source: Revelio Labs workforce data and classified job postings.
Loading Figure 3.7 — Estimates of the change in log employment at AI-adopting firms compared to non-adopters relative to October 2022, estimated separately by seniority level with month fixed effects. Source: Revelio Labs workforce data from professional online profiles.Figure 3.7 — Estimates of the change in log employment at AI-adopting firms compared to non-adopters relative to October 2022, estimated separately by seniority level with month fixed effects. Source: Revelio Labs workforce data from professional online profiles.
Whether AI raises or lowers wages depends on whether productivity complementarity or substitution dominates. We estimate the salary premium associated with AI exposure in activities from job postings, weighting occupations by headcount so the result reflects the labor market as workers experience it rather than as postings are distributed.
Loading Figure 3.8 — The headcount-weighted salary premium per one-standard-deviation increase in a posting's AI exposure. Source: Revelio Labs job postings data.Figure 3.8 — The headcount-weighted salary premium per one-standard-deviation increase in a posting's AI exposure. Source: Revelio Labs job postings data.
AI is changing not just which workers are hired, but what workers do. The mix of activities performed across the economy continues to shift, and most of that change is occurring within occupations rather than through changes in the occupation mix. This distinction matters: job titles may remain stable and people may remain within the same occupations, even as the activities performed inside those jobs change, so analyses focused only on occupation-level employment can miss a substantial part of the transformation.
Loading Figure 4.1 — A taxonomy of Revelio's work-activity classification, sized by appearance in job postings since 2021. A static reference exhibit, not a current-month snapshot. Source: Revelio Labs job postings data.Figure 4.1 — A taxonomy of Revelio's work-activity classification, sized by appearance in job postings since 2021. A static reference exhibit, not a current-month snapshot. Source: Revelio Labs job postings data.
Loading Figure 4.2 — Within- versus between-occupation share of the year-over-year activity dissimilarity index. Source: Revelio Labs workforce data from professional online profiles.Figure 4.2 — Within- versus between-occupation share of the year-over-year activity dissimilarity index. Source: Revelio Labs workforce data from professional online profiles.
Loading Figure 4.3 — Percentage-point change in activity share for the activities gaining and losing the most activity share, 2022 vs. the 2024–2026 average. Source: Revelio Labs job postings data.Figure 4.3 — Percentage-point change in activity share for the activities gaining and losing the most activity share, 2022 vs. the 2024–2026 average. Source: Revelio Labs job postings data.
The impact of adoption on sentiment
We also track whether organizational AI adoption is associated with changes in how employees experience their work. Changing work content shows up in how employees talk about their jobs, not just in postings data. Comparing employee reviews at AI-adopting firms compared to non-adopters isolates how sentiment shifts once a firm actually adopts AI.
Loading Figure 4.4 — Difference-in-differences estimates of firm-level AI adoption on business outlook, job security, and senior-leadership sentiment. Comparisons are made within industry and October 2022 firm-size groups, with each measure standardized by pre-ChatGPT dispersion. Source: Revelio Labs employee reviews, workforce data, and classified job postings.Figure 4.4 — Difference-in-differences estimates of firm-level AI adoption on business outlook, job security, and senior-leadership sentiment. Comparisons are made within industry and October 2022 firm-size groups, with each measure standardized by pre-ChatGPT dispersion. Source: Revelio Labs employee reviews, workforce data, and classified job postings.
Loading Figure 4.5 — The difference in layoff-related language in reviews between AI-adopting and non-adopting firms. Source: Revelio Labs sentiment data from employee reviews.Figure 4.5 — The difference in layoff-related language in reviews between AI-adopting and non-adopting firms. Source: Revelio Labs sentiment data from employee reviews.
The hiring process is producing fewer successful matches. Job postings per external hire have risen for six years, while candidates increasingly report recruiter ghosting and poor communication. These trends predate ChatGPT, so they cannot be attributed entirely to generative AI. But AI may be making an existing problem worse by enabling candidates to submit more polished applications, weakening the signals employers use to identify strong candidates.
Job Postings Are Producing Fewer Hires
Loading Figure 5.1 — Job postings per external hire over time, shown as a trailing 12-month ratio. Postings are matched to external hires at the same company, state, role, and seniority level over the posting month and following two months. Source: Revelio Labs workforce and job-posting data.Figure 5.1 — Job postings per external hire over time, shown as a trailing 12-month ratio. Postings are matched to external hires at the same company, state, role, and seniority level over the posting month and following two months. Source: Revelio Labs workforce and job-posting data.
Recruiter Ghosting Is Back Near Its October 2022 Baseline
Loading Figure 5.2 — Mentions of recruiter ghosting in no-offer interview reviews over time. Source: Revelio Labs interview-review data.Figure 5.2 — Mentions of recruiter ghosting in no-offer interview reviews over time. Source: Revelio Labs interview-review data.
Poor Communication Is Becoming a Larger Source of Candidate Dissatisfaction
Loading Figure 5.3 — Share of no-offer interview complaints mentioning communication, process speed, and other hiring issues. Source: Revelio Labs interview-review data.Figure 5.3 — Share of no-offer interview complaints mentioning communication, process speed, and other hiring issues. Source: Revelio Labs interview-review data.
Try Revelio Labs
Start understanding the world's workforce.