Skip to content
ContentLora

    Tip: press / anywhere to search.

    Analysis

    Is AI taking jobs? What the labour data shows in 2026

    A Stanford working paper (not peer reviewed) finds no broad, economy-wide job loss from generative AI, but a growing gap for young workers in the most exposed occupations.[1][2] A Stanford working paper, not yet peer reviewed, using ADP payroll data through June 2026 found employment of 22- to 25-year-olds in AI-exposed jobs 19% below the path of less-exposed peers, mainly through less hiring.[2][3] Other analyses find little disruption to the wider labour market, and forecasts of future displacement vary widely.[4][5][6]

    Editor reviewedUpdated Analysis17 claims · 53% primaryArtificial intelligence, Economy and trade
    Show:

    As of November 2026, the main labour-market studies show no broad job loss from generative AI, but a widening shortfall in employment for young workers in the most exposed occupations.[1][2] The debate is over whether that gap is an early warning or a narrow effect, and how much more change is coming.[7][6]

    This page sets out what the evidence shows and the positions of named researchers and institutions. Interpretation and forecasts are labelled. Company usage data are self-reported, and working papers have not completed peer review.

    What payroll data show

    An August 2026 update of a Stanford Digital Economy Lab working paper, which has not been peer reviewed, uses ADP payroll records covering millions of US workers through June 2026. It found no evidence of widespread, economy-wide job displacement.[1] Employment of workers aged 22 to 25 in AI-exposed occupations stood 19% below where it would have been had it kept pace with less-exposed peers. Experienced workers showed no comparable gap, and the gap has widened since the authors first reported it in August 2025.[2]

    The authors say the gap comes mainly from less hiring of young workers rather than more separations. It is concentrated in occupations where AI mainly substitutes for human tasks.[3] They note that the patterns weaken when controlling for education, show some trends that began before generative AI, and are stronger in the ADP sample than in national surveys. They call them early indicators rather than causal estimates.[7]

    What broader measures and surveys show

    The Budget Lab at Yale reported in October 2025 that its measures showed no discernible disruption to the broader labour market in the 33 months since ChatGPT’s release.[4] It found the mix of occupations changing somewhat faster than in the past, but said the difference was small and began before AI was widely used at work.[8]

    The New York Fed’s August 2025 regional business surveys found a notable rise in AI use but very few AI-induced layoffs. Retraining was more likely than job loss, though some firms were scaling back hiring, and firms expected AI-related layoffs and hiring cuts to increase.[5] Overall, US payrolls rose by 165,000 in October 2026 and unemployment was 4.1%.[9]

    Read together, the payroll and survey evidence is consistent: firms are adjusting mainly by hiring fewer entry-level workers in exposed roles, not by laying off existing staff. That would show up first among young workers and only later, if at all, in the headline unemployment rate.

    How AI is being used at work

    Anthropic publishes an Economic Index based on its own Claude usage data; the figures below are self-reported by the company. Its September 2025 report found that “directive” conversations, in which users hand over complete tasks, rose from 27% to 39%.[10] It also found that 77% of business uses of its API showed automation patterns, compared with about 50% for Claude.ai users.[11] In a June 2026 survey of Claude users, which the company says is not representative of the population, early-career workers said AI could do the largest share of their work and were most worried about job loss.[12]

    Official projections and forecasts

    In February 2025 the Bureau of Labor Statistics projected software developer jobs to grow 17.9% from 2023 to 2033. It projected declines of 4.7% for medical transcriptionists and 5.0% for customer service representatives, occupations it expects AI to affect.[13] BLS said the timing and scale of many generative AI effects are too uncertain to include in its projections.[14] Goldman Sachs Research, the bank’s analyst team, estimated in August 2025 that 6-7% of the US workforce could be displaced if AI is widely adopted. It expects the effect to be transitory, with unemployment about half a percentage point higher during the transition.[6]

    The forecasts differ mostly in their assumptions about adoption speed and how quickly displaced workers find new roles. None of them is yet tested against data from a period of widespread use of AI agents.

    Where the debate goes next

    The Stanford authors publish their series as public AI Economic Indicators for ongoing tracking, so the young-worker gap can be checked as new payroll data arrive.[15] BLS’s own projections leave many generative AI effects out until the evidence is clearer.[14]

    Related reading: the AI copyright lawsuits tracker covers the legal fight over training data, and the US state AI laws tracker covers new rules on automated decision-making.

    Colorado’s revised AI law, for example, regulates automated decision-making technology used in consequential decisions from 2027, as tracked in the US state AI laws tracker.[16] The AI copyright lawsuits tracker follows cases such as the $1.5 billion settlement between Anthropic and book authors.[17]

    Competing views

    Early warning for young workers

    Payroll data show a widening employment gap for 22- to 25-year-olds in AI-exposed occupations, driven by reduced hiring and concentrated where AI substitutes for tasks. The authors call these early indicators, not causal estimates.[2][3][7]

    No broad disruption yet

    Measures of the overall occupational mix show no discernible disruption since ChatGPT's release, and firms surveyed by the New York Fed reported very few AI-induced layoffs, favouring retraining.[4][8][5][1]

    Bigger effects ahead

    Usage data show AI increasingly being handed whole tasks, firms expect AI-related layoffs and hiring cuts to rise, and Goldman Sachs Research estimates 6-7% of US workers could be displaced if AI is widely adopted, though it expects the effect to be transitory.[10][11][5][6]

    You've reached the end of the course. Next crash course

    Internet and telecoms: a crash course

    A sourced crash course on the internet, DNS, Wi-Fi, fibre, 5G and satellites, who governs them, and US rules on spectrum, net neutrality and robocalls.

    More in AI

    Questions readers ask

    Is AI causing unemployment to rise?

    The studies cited here do not show broad job loss so far. A Stanford working paper using ADP payroll data through June 2026 found no widespread, economy-wide displacement, and the Budget Lab at Yale found no discernible disruption to the broader labour market in the 33 months after ChatGPT's release. US unemployment was 4.1% in October 2026.[1][4][9]

    Are young workers being hit hardest?

    In the Stanford data, employment of workers aged 22 to 25 in AI-exposed occupations was 19% below where it would have been had it kept pace with less-exposed peers, with no comparable gap for experienced workers. The authors say this is descriptive, not proof that AI caused it.[2][7]

    Are companies laying people off because of AI?

    Few, according to the New York Fed's August 2025 regional surveys. Firms reported rising AI use but very few AI-induced layoffs, and were more likely to retrain staff. Some were scaling back hiring, and firms expected AI-related layoffs and hiring cuts to increase.[5]

    Which jobs does the government expect AI to shrink?

    In February 2025, BLS projected declines through 2033 of 4.7% for medical transcriptionists and 5.0% for customer service representatives, and growth of 17.9% for software developers. It said the timing and scale of many generative AI effects are too uncertain to include.[13][14]

    How many jobs could AI eventually replace?

    Forecasts vary. Goldman Sachs Research, the bank's analysts, estimated in August 2025 that 6-7% of the US workforce could be displaced if AI is widely adopted, but expected the effect to be transitory, with unemployment about half a point higher during the transition.[6]

    Sources

    · 17 claims from 10 publishers

    Each numbered claim is a statement we checked against the sources listed with it. Tap a number in the text to see its sources in place.

    1. [1]

      An August 2026 update of a Stanford working paper (not peer reviewed) using ADP payroll data on millions of US workers through June 2026 found no evidence of widespread, economy-wide job displacement from generative AI. reportedas of 2026-08-01

    2. [2]

      The same paper found that employment of workers aged 22 to 25 in AI-exposed occupations was 19% below where it would have been had it kept pace with less-exposed peers, with no comparable gap for experienced workers, and that the gap had widened since August 2025. reportedas of 2026-08-01

    3. [3]

      The authors report that the divergence works mainly through less hiring of young workers rather than more separations, and is concentrated in occupations where AI mainly substitutes for human tasks. reportedas of 2026-08-01

    4. [4]

      The Budget Lab at Yale reported on 1 October 2025 that its measures showed the broader US labor market had not experienced a discernible disruption in the 33 months since ChatGPT's release. reportedas of 2025-10-01

    5. [5]

      In August 2025 regional business surveys, the New York Fed found a notable rise in firms' AI use but very few AI-induced layoffs, with retraining more likely than job loss; some firms were scaling back hiring because of AI, and firms expected AI-related layoffs and hiring cuts to increase. confirmedas of 2025-09-04

    6. [6]

      Goldman Sachs Research estimated in August 2025 that AI could displace 6-7% of the US workforce if widely adopted, but expected the effect to be transitory, with unemployment rising by about half a percentage point during the transition. reportedas of 2025-08-13· forecast

    7. [7]

      The authors say the patterns weaken when controlling for education, show some divergent trends before generative AI, and are stronger in the ADP sample than in national survey benchmarks; they describe them as early descriptive indicators, not causal estimates. reportedas of 2026-08-01

    8. [8]

      The Yale analysis found the occupational mix changing somewhat faster than in the past, but said the difference was not large and began before AI was widely used at work. reportedas of 2025-10-01

    9. [9]

      US nonfarm payrolls rose by 165,000 in October 2026 and the unemployment rate was little changed at 4.1%. confirmedas of 2026-11-06

    10. [10]

      Anthropic's September 2025 Economic Index report, based on its own Claude usage data, found the share of "directive" conversations, in which users delegate complete tasks, rose from 27% to 39%. confirmedas of 2025-09-15

    11. [11]

      The same report found that 77% of business uses of Anthropic's first-party API showed automation patterns, compared with about 50% for Claude.ai users. confirmedas of 2025-09-15

    12. [12]

      In Anthropic's June 2026 Economic Index survey of Claude users, which it says is not representative of the general population, early-career workers reported that AI can do the largest share of their work and expressed the most concern about job loss. confirmedas of 2026-06-26

    13. [13]

      In February 2025 the Bureau of Labor Statistics projected software developer employment to grow 17.9% from 2023 to 2033, while projecting declines of 4.7% for medical transcriptionists and 5.0% for customer service representatives, occupations it expects AI to affect. confirmedas of 2025-02-01· forecast

    14. [14]

      BLS said the timing and scale of many potential impacts of generative AI are too uncertain to be reflected in its projections. confirmedas of 2025-02-01

    15. [15]

      The Stanford authors provide a public set of AI Economic Indicators to support ongoing tracking of AI's effects on the economy. reportedas of 2026-08-01

    16. [16]

      Colorado's SB26-189, signed by the governor on 14 May 2026, revised the 2024 law, repealing and re-enacting it with new requirements on automated decision-making technology (ADMT) that influences consequential decisions, beginning in 2027. confirmedas of 2026-05-14

    17. [17]

      The Bartz v. Anthropic class settlement created a non-reversionary fund of $1.5 billion covering 482,460 works, an estimated payment of about $3,000 per work. confirmedas of 2026-07-20

    Revision history (2)
    1. The summary now credits its finding to the one Stanford working paper it cites, not to the main studies in general.
    2. Page created.

    Created Nov 7, 2026. Last reviewed by an editor on Nov 8, 2026. Next scheduled review: Feb 7, 2027.

    Cite this page

    "Is AI taking jobs? What the labour data shows in 2026." ContentLora, updated Nov 14, 2026. https://contentlora.com/analysis/ai-and-jobs-evidence-debate

    Spotted an error? Suggest a correction or emailcorrections@contentlora.com.