AI Job Displacement Statistics 2026
By Axis Intelligence Research
Co-author: David Park | Last updated: October 29, 2026 | License: CC BY 4.0
U.S. employers attributed 120,136 announced job cuts to artificial intelligence from January to September 2026, according to Challenger, Gray & Christmas. Axis Intelligence Research finds those AI-cited cuts equal 0.84% of all U.S. layoffs and discharges the Bureau of Labor Statistics recorded from January to August 2026: about one layoff in 119.
Quick Answer: How Many Jobs Has AI Displaced in 2026?
AI-cited job cuts reached 120,136 through September 2026, more than double the 54,836 for all of 2025 (Challenger). That is still under 1% of all U.S. layoffs. The sharper displacement signal is at the entry level: workers aged 22 to 25 in AI-exposed jobs are 19% below less-exposed peers (Stanford, August 2026).
Key Findings
- Axis Intelligence Research finds that AI-cited announced job cuts equaled 0.84% of U.S. layoffs and discharges in January–August 2026, up from 0.26% in 2025, based on Challenger and BLS JOLTS data.
- Challenger, Gray & Christmas counted 120,136 AI-cited U.S. job cuts from January through September 2026, about 21% of all announced cuts, after 54,836 in the whole of 2025.
- Axis Intelligence Research finds that the unemployment rate of recent U.S. college graduates has exceeded the rate for all workers for 66 consecutive months through June 2026, the longest run in New York Fed data going back to 1990.
- Employment of U.S. workers aged 22–25 in AI-exposed occupations was 19% below less-exposed peers through June 2026, per the Stanford Digital Economy Lab’s revised ADP payroll analysis.
- Axis Intelligence Research estimates that AI-cited cuts in 2026 are running at 11% to 17% of the yearly displacement pace implied by Goldman Sachs’s estimates of 6% to 9% of U.S. workers displaced over a decade.
How Many Jobs Has AI Eliminated in 2026? The Month-by-Month Count
The only running U.S. count of layoffs that employers themselves tie to AI comes from the outplacement firm Challenger, Gray & Christmas, which began tracking AI as a stated reason in 2023. In its September 2026 report, AI remained the leading reason year to date, cited in 120,136 announcements, about 21% of the 573,195 total.
The year was not a straight line. AI was cited in 38,579 cuts in May alone, which Challenger called the highest monthly total since it began tracking the reason. Three months later the count had fallen by roughly 91%.
AI-cited U.S. job cuts by month, 2026
| Period | AI-cited cuts | Source |
|---|---|---|
| January–February | 12,304 | Challenger (YTD through March minus March) |
| March | 15,341 | Challenger, March 2026 |
| April | 21,490 | Challenger (YTD through May minus May and Q1) |
| May | 38,579 | Challenger, May 2026 |
| June | 14,029 | Challenger, June 2026 |
| July | 10,970 | Challenger, July 2026 |
| August | 3,462 | Challenger, August 2026 |
| September | 3,961 | Challenger, September 2026 |
| January–September | 120,136 | Challenger, September 2026 |
January–February and April are Axis Intelligence Research subtractions from Challenger’s year-to-date totals; Challenger does not publish them as standalone months. The nine rows sum exactly to Challenger’s 120,136.
For scale, Challenger counted 54,836 AI-cited cuts in all of 2025, out of 1,206,374 total announcements, and 71,825 since it began tracking AI in 2023. Adding 2026 brings the cumulative AI-cited count to 191,961 as of September 30, 2026.
David Park’s read: the May spike is the number that made headlines, and the August–September floor is the one that should shape a 2027 plan. A count that swings elevenfold in three months is measuring how companies describe cuts as much as how many they make. Challenger itself listed AI as only the fourth-most cited reason in August, behind categories that a company could use for the same restructuring. The company-by-company evidence on who says what, and in which document, lives on our AI layoff tracker.
What Share of All U.S. Layoffs Is Caused by AI? The AI Displacement Visibility Ratio (ADVR)
Headline AI layoff counts are rarely set against the size of the labor market they sit in. The AI Displacement Visibility Ratio (ADVR) is the Axis Intelligence Research metric that does it: the share of all U.S. layoffs and discharges that employers publicly attribute to AI.
Formula:
ADVR = AI-cited announced job cuts (Challenger) ÷ total layoffs and discharges (BLS Job Openings and Labor Turnover Survey, JOLTS) × 100, for the same period.
Layoffs and discharges are involuntary separations initiated by the employer, as counted by JOLTS. The ratio is called a visibility ratio because the numerator captures only the displacement a company chooses to attribute to AI in public.
ADVR readings, 2025 and 2026
| Period | AI-cited cuts (Challenger) | Layoffs and discharges (BLS JOLTS) | ADVR | About one layoff in… |
|---|---|---|---|---|
| Full year 2025 | 54,836 | 21.2 million | 0.26% | 387 |
| January–February 2026 | 12,304 | 3,374,000 | 0.36% | 274 |
| March 2026 | 15,341 | 1,884,000 | 0.81% | 123 |
| April 2026 | 21,490 | 1,667,000 | 1.29% | 78 |
| May 2026 | 38,579 | 1,761,000 | 2.19% | 46 |
| June 2026 | 14,029 | 1,785,000 | 0.79% | 127 |
| July 2026 | 10,970 | 1,702,000 | 0.64% | 155 |
| August 2026 | 3,462 | 1,641,000 (preliminary) | 0.21% | 474 |
| January–August 2026 | 116,175 | 13,814,000 | 0.84% | 119 |
Sources: Challenger, Gray & Christmas monthly reports; BLS JOLTS releases of March 13, June 2, June 30, August 4 and September 29, 2026. Monthly JOLTS values are seasonally adjusted and taken from the latest published vintage; the 2025 annual figure is not seasonally adjusted. ADVR calculated by Axis Intelligence Research.
Axis Intelligence Research finds the ADVR more than tripled between 2025 and January–August 2026, from 0.26% to 0.84%. Even at its May peak, AI-cited announcements covered about 1 in 46 U.S. layoffs.
A second check against the BLS displaced-worker count
The BLS counted 7.4 million workers displaced from January 2023 to December 2025. Over the same three years, Challenger counted 71,825 AI-cited cuts. Axis Intelligence Research calculates that AI-cited announcements equaled 0.97% of that displaced-worker total. Two different federal denominators, one survey and one establishment-based, land on the same answer: under 1%.
Neither measure counts AI’s whole effect. JOLTS and the displaced-worker survey record separations that happened; they do not ask why. Challenger records announcements, which can arrive months before or after the separation. The ADVR therefore measures attributed displacement against all displacement, not AI’s true causal share.
David Park’s read: for anyone lying awake after a town hall that used the word “efficiency,” the ratio is the honest denominator. Ninety-nine of every hundred people laid off in the United States this year lost their job for a reason their employer did not call AI. That does not make your role safe. It means the risk shows up less as a mass layoff and more as a requisition that never reopens, which is the next section.
Why Isn’t AI Displacement Showing Up in the Unemployment Rate?
The U.S. unemployment rate was 4.2% in September 2026, with 7.1 million people unemployed, per the BLS. Payrolls rose by just 29,000 that month, and the long-term unemployed made up 27.1% of all unemployed people.
The labor market is slow, but the slowness runs through hiring, not firing. In August 2026 the JOLTS hires rate was 3.3% and the layoffs and discharges rate was 1.0%. A low layoff rate with a low hiring rate is the pattern economists call low-hire, low-fire: people who keep their jobs are mostly fine, and people looking for one wait longer.
The best causal tests so far find no measurable AI effect on overall employment. A May 7, 2026 analysis by The Budget Lab at Yale estimated that AI’s effect on the employment share of the average AI-exposed occupation is close to zero and not statistically distinguishable from zero, and found the same for wages. Yale also noted that early-2026 hiring was especially weak for unemployed job seekers.
Goldman Sachs economists reached a larger number with a different tool. Their U.S. Daily note by economist Elsie Peng, as reported by Fortune on April 6, 2026, estimated that AI erased roughly 16,000 net U.S. jobs a month over the prior year: about 25,000 lost to substitution, offset by about 9,000 added through augmentation. Fortune reported that the estimate comes from regressions on exposure scores combined with an IMF complementarity index, not a count of actual job losses, and that Goldman’s economists themselves cautioned the true aggregate effect is likely smaller because the model does not fully capture hiring tied to data centers and power infrastructure.
Three AI job-loss numbers, three different things
| Figure | What it measures | Source |
|---|---|---|
| 120,136 (Jan–Sep 2026) | Layoff announcements where the employer named AI | Challenger, Gray & Christmas |
| ~16,000 net per month (prior year to April 2026) | Regression estimate of AI’s net effect on payrolls | Goldman Sachs note, as reported by Fortune |
| Close to zero, not statistically significant | Causal estimate of AI’s effect on exposed occupations’ employment share | The Budget Lab at Yale, May 2026 |
The numbers do not contradict each other as much as they appear to. A modeled loss of 16,000 a month is roughly 192,000 a year, larger than the 54,836 Challenger counted for 2025, because a regression can pick up hiring that never happened, while an announcement count cannot. That gap is the displacement nobody announces.
Who Is Being Displaced by AI? Entry-Level Workers First
The clearest AI displacement signal in U.S. data is age. The Stanford Digital Economy Lab’s August 12, 2026 revision of its “Canaries in the Coal Mine” paper, using ADP payroll records through June 2026, finds that employment of workers aged 22 to 25 in AI-exposed occupations is 19% below where it would be had it kept pace with less-exposed peers. The authors report no comparable gap for experienced workers.
The gap has widened. The first version, published in August 2025, reported a 13% relative decline. The revision also locates the mechanism: the gap comes mainly from reduced hiring of young workers, not from more separations, and it is concentrated where AI substitutes for tasks rather than complements them. The authors describe their results as descriptive, not causal.
Graduates feel it first. The Federal Reserve Bank of New York reports that recent college graduates (ages 22 to 27) had an unemployment rate of about 5.6% in the second quarter of 2026, and that 42% were underemployed in jobs that do not typically require a degree.
The graduate penalty: 66 months and counting
A college degree used to lower the odds of unemployment right out of school. That is no longer true. In the New York Fed’s monthly series, the unemployment rate for recent graduates was 5.698% in June 2026, against 4.061% for all workers: a gap of 1.64 percentage points.
Axis Intelligence Research finds that recent graduates’ unemployment has been higher than the all-worker rate for 66 consecutive months, from January 2021 through June 2026, the longest such run since the New York Fed series begins in 1990. The previous longest run was 11 months, from May 2019 to March 2020. From 1990 to 2019, recent graduates’ rate averaged 1.29 points below the all-worker rate. The widest gap on record came in April 2025, at 1.78 points; June 2026 is within 0.14 points of it.
| Indicator (New York Fed series) | Value | As of |
|---|---|---|
| Recent graduates’ unemployment rate | 5.698% | June 2026 |
| All workers’ unemployment rate | 4.061% | June 2026 |
| Gap, recent graduates minus all workers | +1.64 points | June 2026 |
| Average gap, 1990–2019 | −1.29 points | Axis calculation |
| Consecutive months with recent graduates above all workers | 66 | January 2021–June 2026 |
| Widest gap in the series | +1.78 points | April 2025 |
Source: Federal Reserve Bank of New York, The Labor Market for Recent College Graduates, data file updated August 6, 2026; streak and averages calculated by Axis Intelligence Research.
The majors closest to AI’s strongest use case are not protected. In the New York Fed’s outcomes-by-major table (updated February 4, 2026), computer engineering graduates had a 7.8% unemployment rate and computer science graduates 7.0%, the second- and fourth-highest of 73 majors, against 4.2% for all majors. Both still carry the highest early-career pay of the group, at $90,000 and $87,000.
The most AI-exposed industry is shrinking. Information-sector payrolls fell from 2,859,000 in September 2025 to 2,739,000 in September 2026 (preliminary), per BLS Table B-1. Axis Intelligence Research calculates that is a 4.2% drop in a year, while professional and business services grew 0.55%. The Information sector also includes publishing, broadcasting and telecommunications, so not all of that decline is AI.
Entry-level displacement signals, 2026
| Indicator | Value | As of | Source |
|---|---|---|---|
| Employment gap, ages 22–25, AI-exposed vs less-exposed occupations | −19% | June 2026 | Stanford Digital Economy Lab (ADP data) |
| Recent college graduate unemployment rate (ages 22–27) | 5.6% | Q2 2026 | Federal Reserve Bank of New York |
| Recent college graduate underemployment rate | 42% | Q2 2026 | Federal Reserve Bank of New York |
| Months recent graduates’ unemployment has exceeded all workers’ | 66 in a row | June 2026 | Axis calculation from New York Fed data |
| Information-sector employment change, year over year | −4.2% | September 2026 | Axis calculation from BLS Table B-1 |
| Organizations expecting AI to shrink their workforce next year | 32% | 2026 | Stanford HAI AI Index 2026 |
Young people have noticed. In a Gallup survey of 1,572 Americans aged 14 to 29, fielded February 24 to March 4, 2026 with the Walton Family Foundation, 31% said AI makes them feel angry, up 9 points from 2025, while 22% said excited, down 14 points. Among employed Gen Z respondents, 48% said AI’s workplace risks outweigh its benefits, up from 37%.
David Park’s read: the entry rung has not been sawn off. It has been raised. Fewer junior requisitions open, and the ones that do ask for judgment that used to be learned on the job. For a 2027 graduate, the practical answer is to target roles where AI complements the work rather than substitutes for it, because that is where Stanford finds employment holding up. Our Gen Z employment statistics track the age gap month by month, and the AI job market statistics show where AI-skill postings are growing.
What Happens to Workers After They Are Displaced?
The BLS Displaced Worker Survey is the official record of what happens after a job loss. It does not ask whether AI was involved, but it shows how hard landing has become as AI spreads through the occupations that lost the most workers.
According to the BLS release of August 27, 2026, 7.4 million workers were displaced from January 2023 to December 2025, up from 6.3 million in the 2021–2023 survey. Of these, 3.3 million were long-tenured workers who had held their job for three years or more.
Long-tenured displaced workers, 2023–2025: where they were in January 2026
| Measure | January 2026 | January 2024 (prior survey) |
|---|---|---|
| Reemployed | 66.1% | 65.7% |
| Unemployed | 18.3% | 16.1% |
| Reemployed full-time workers earning as much or more than before | about 49% | about 62% |
| Job lost because position or shift was abolished | 44.4% | — |
| Unemployed, displaced from professional and business services | 27.1% | — |
| Reemployed, displaced from the information industry | 76.8% | — |
| Unemployed, displaced from professional and related occupations | 20.4% | — |
| Reemployed, ages 55–64 | 57.3% | — |
Source: BLS, Worker Displacement: 2023–25 (USDL-26-1423), August 27, 2026.
Two numbers matter most. About two in three long-tenured displaced workers found new jobs, but only about half of the reemployed full-time workers matched their old pay, down from roughly six in ten in the previous survey. The pay cut, not the job loss, is the larger change.
“Position or shift abolished” was the most common reason, at 44.4%. That is the category an AI-driven reorganization would fall into, though the BLS does not separate AI from any other cause. Displaced professional and business services workers, a group full of the analyst, support and coordination roles that generative AI touches, had an unemployment share of 27.1%, double that of displaced manufacturing workers (13.5%) and more than three times that of displaced information-industry workers (8.1%).
David Park’s read: if you are already in a severance conversation, the published median is your anchor. Two-thirds land somewhere; half of those land at the same pay or better. Negotiate severance as if the search will take longer and pay less than your last offer, because the federal data says it often does.
How Many Jobs Could AI Displace by 2027 and 2030?
The long-range estimates measure potential, intentions or modeled transitions, not jobs lost. Each one answers a different question.
The major AI displacement projections, side by side
| Projection | Figure | What it measures | Source, date |
|---|---|---|---|
| U.S. workers displaced over the AI transition (latest) | Up to about 9% over a decade | Modeled displacement, U.S. and other developed economies | Goldman Sachs (Joseph Briggs), as reported by Fortune, October 22, 2026 |
| U.S. workers displaced over the AI transition (earlier base case) | 6–7% over about 10 years (range 3–14%) | Base-case modeled displacement if AI is widely adopted | Goldman Sachs Research, March 18, 2026 |
| Unemployment rate effect | +0.6 percentage points | If the transition takes a decade | Goldman Sachs Research, March 2026 |
| U.S. work hours technically automatable | 57% (44% by agents, 13% by robots) | Technical potential, “not a forecast of job losses” | McKinsey Global Institute, November 25, 2025 |
| Global jobs created / displaced, 2025–2030 | 170 million / 92 million (net +78 million) | Employer survey, 55 economies, 14.1 million workers | WEF Future of Jobs Report 2025, January 7, 2025 |
| Employers expecting to cut staff where AI automates tasks | 40% | Employer intention | WEF Future of Jobs Report 2025 |
| Global employment exposed to generative AI | 25% (34% in high-income countries) | Task exposure | ILO–NASK Global Index, May 20, 2025 |
The WEF, McKinsey and ILO figures are more than six months old; no newer editions have been published.
Goldman raised its own estimate during 2026. Fortune reported on October 22, 2026 that Goldman economist Joseph Briggs now puts displacement at up to about 9% of workers, up from the 6–7% (within a 3–14% range) he and Sarah Dong estimated in August 2025, with the same assumption that AI eventually lifts productivity by 15%. Briggs expects the annual rise in unemployment to stay under one percentage point even at 9%.
The ILO adds the demographic edge most coverage misses: in high-income countries, jobs in the highest automation-risk category make up 9.6% of female employment against 3.5% of male employment. The WEF also expects 39% of workers’ core skills to change by 2030, and 85% of employers say they will prioritize upskilling.
How far is visible AI displacement from Goldman’s estimates?
Axis Intelligence Research set the observed 2026 pace against Goldman’s estimates. Applied to the 159,044,000 nonfarm payroll jobs the BLS counted in September 2026 and spread evenly over ten years, Goldman’s 6% implies 954,264 displaced workers a year, 7% implies 1,113,308, and the latest 9% implies 1,431,396. AI-cited announcements are running at 160,181 a year (120,136 for January–September × 12 ÷ 9).
Axis Intelligence Research estimates that visible, employer-attributed AI displacement is running at 11% to 17% of the yearly pace Goldman’s estimates imply. Either the transition has barely started, or most of it is arriving as unannounced hiring cuts. The Stanford hiring data points to the second.
Axis 2027 run-rate scenarios for AI-cited job cuts
These are run-rate scenarios, not forecasts. Each repeats a pace already observed in 2026 for twelve months and divides it by 2025’s 21.2 million layoffs and discharges to give the implied ADVR.
| Scenario | 2026 pace repeated | Monthly average | AI-cited cuts in 2027 | Implied ADVR |
|---|---|---|---|---|
| Cooling | August–September 2026 | 3,712 | 44,538 | 0.21% |
| Base | January–September 2026 | 13,348 | 160,181 | 0.76% |
| Surge | March–July 2026 | 20,082 | 240,982 | 1.14% |
On the same arithmetic, Axis Intelligence Research brackets the 2026 full-year AI-cited count between 131,271 (if the fourth quarter runs at the August–September pace) and 160,181 (if it runs at the January–September average). Even the surge scenario keeps AI-cited cuts near 1 in 88 U.S. layoffs.
What changes in 2027
On December 2, 2027, EU AI Act rules for high-risk AI systems in employment, covering recruitment, worker management and evaluation, begin to apply, per the European Commission. Emotion recognition in workplaces has been banned since February 2025. Any employer using AI to screen candidates or assess staff in the EU will have to meet the Act’s high-risk requirements, and that compliance paperwork may make AI’s role in workforce decisions more visible than any U.S. disclosure does today. Deadlines are tracked on our EU AI Act compliance tracker, and the wider shifts in how work is organized are projected in our future of work trends for 2027.
The bigger 2027 variable may be the business cycle. In the same October 2026 work, Briggs found that technology-driven job losses stay limited in healthy economies and accelerate in downturns, when companies cut labor costs; the share of Americans in routine jobs held steady in expansions and fell sharply in the recessions of the early 1980s, 2001, 2008 and 2020. Fortune also reported Goldman’s estimate that only 23% of U.S. establishments have built AI into regular workflows. A 2027 slowdown would therefore be the most likely trigger for the surge scenario.
The signposts that will move the ADVR in 2027: Challenger’s monthly reports, the monthly JOLTS release, the New York Fed’s quarterly graduate data, and the next update of the Stanford ADP series. A 2027 ADVR above 1.14% for the full year would mean the surge pace has become the normal pace.
David Park’s read: plan for the base case and price in the surge. For workers, that means building the skills listed in job ads now, not after a reorganization; for HR teams setting 2027 comp bands and requisition budgets, it means treating “AI-attributed” headcount moves as a disclosure question, not only a cost question.
Methodology
What this page measures. Displacement here means involuntary job loss, counted three ways: employer announcements that name AI (Challenger), all employer-initiated separations (BLS JOLTS layoffs and discharges), and survey-reported displacement of workers (BLS Displaced Worker Survey). The entry-level analysis adds payroll employment by age (Stanford/ADP) and graduate outcomes (New York Fed). The graduate-penalty streak and averages use the New York Fed’s monthly data file for recent graduates and all workers, 1990 to June 2026.
How the ADVR is built. ADVR = Challenger AI-cited cuts ÷ JOLTS layoffs and discharges × 100 for matching periods. Monthly JOLTS levels are seasonally adjusted, from BLS series JTS000000000000000LDL as retrieved on October 29, 2026; they match the figures in the monthly releases linked above. The 2025 reading uses the JOLTS annual total of 21.2 million, which is not seasonally adjusted; the sum of the twelve seasonally adjusted 2025 months (21,169,000) gives the same 0.26%. Challenger figures are not seasonally adjusted. Monthly Challenger values for January–February and April are derived by subtraction from published year-to-date totals.
Why the ratio is read as a lower bound. Announcements name AI only when a company chooses to, an announcement can precede the separation by months, and JOLTS counts every layoff, including those with no announcement. The ADVR shows how much displacement is publicly attributed to AI; it does not estimate how much AI caused.
Scenarios. The 2027 run-rate scenarios repeat an observed 2026 monthly average for twelve months and hold the JOLTS denominator at its 2025 level. The Goldman comparison divides the 6%, 7% and 9% estimates evenly over ten years and applies it to September 2026 nonfarm payrolls; Goldman does not publish a year-by-year path.
What we left out. Company-level layoff events are not re-counted here; they are classified by evidence tier on the AI layoff tracker. Exposure estimates in the hundreds of millions are reconciled on our AI job loss statistics page, sector-level confirmed cuts on AI replacing jobs statistics, and technology-sector layoff totals on tech layoffs statistics. Older figures are flagged in the CSV.
About This Dataset
The file ai-job-displacement-statistics.csv holds every number on this page in long format: Challenger AI-cited cuts by month for 2026 and annual totals since 2023, JOLTS layoffs and discharges by month, every ADVR reading with its formula, the BLS Displaced Worker Survey outcomes, the entry-level indicators, the long-range projections and the 2027 scenarios. Each row carries source_org, source_document, source_url, retrieved_date (October 29, 2026), is_primary, axis_calculated and a method_note for every Axis calculation.
Cite This Research
APA: Axis Intelligence Research & Park, D. (2026, October 29). AI job displacement statistics 2026: Under 1% of U.S. layoffs cite AI, and the entry rung is where it shows [Dataset]. Axis Intelligence. https://axis-intelligence.com/ai-job-displacement-statistics/
MLA: Axis Intelligence Research, and David Park. “AI Job Displacement Statistics 2026: Under 1% of U.S. Layoffs Cite AI, and the Entry Rung Is Where It Shows.” Axis Intelligence, 29 Oct. 2026, axis-intelligence.com/ai-job-displacement-statistics/.
Chicago: Axis Intelligence Research, and David Park. “AI Job Displacement Statistics 2026: Under 1% of U.S. Layoffs Cite AI, and the Entry Rung Is Where It Shows.” Axis Intelligence, October 29, 2026. https://axis-intelligence.com/ai-job-displacement-statistics/.
AI Job Displacement: Questions People Ask
Does the BLS count jobs lost to AI?
No. The BLS records layoffs and the reasons workers give for losing a job, such as a plant closing, insufficient work or a position being abolished, but it has no AI category. The only U.S. count of layoffs attributed to AI by employers is Challenger, Gray & Christmas, at 120,136 for January–September 2026.
Why do reports on AI job losses disagree so much?
They count different things. Challenger counts announcements that name AI (54,836 in 2025). Goldman Sachs economists modeled about 16,000 net jobs a month lost to AI over the year to April 2026, which captures hiring that never happened. Yale’s Budget Lab found no statistically significant effect on employment in exposed occupations. A figure in the hundreds of millions is almost always task exposure, not job loss.
Are companies blaming AI for layoffs they would have made anyway?
The data cannot settle it, but it points both ways. AI-cited cuts fell from 38,579 in May 2026 to 3,462 in August, when AI dropped to fourth among Challenger’s stated reasons. Either fewer cuts were AI-driven, or companies stopped saying so. Our AI layoff tracker separates cuts named in filings from those attributed only by the press.
If I lose my job, how likely am I to find a new one, and at what pay?
Of long-tenured workers displaced in 2023–2025, 66.1% were reemployed by January 2026, per the BLS. About 49% of those back in full-time jobs earned as much or more than before, down from about 62% two years earlier. Workers displaced from professional and business services had a 27.1% unemployment share, against 13.5% for manufacturing.
Is the class of 2027 graduating into an AI hiring freeze?
The hiring door is narrower, not shut. Employment of 22- to 25-year-olds in AI-exposed occupations is 19% below less-exposed peers (Stanford, June 2026 data), driven by fewer hires rather than more layoffs. Recent graduates faced 5.6% unemployment and 42% underemployment in Q2 2026, per the New York Fed, and their unemployment rate has topped the all-worker rate for 66 straight months, a record in data back to 1990.
What happens to AI in hiring on December 2, 2027?
EU AI Act rules for high-risk AI systems in employment, including recruitment and worker management, start to apply, per the European Commission. Employers using AI to screen or rank candidates in the EU must meet the Act’s high-risk obligations from that date. Emotion recognition in workplaces has been banned since February 2025.
How many U.S. jobs could AI displace over the next decade?
Goldman Sachs’s March 2026 base case was 6–7% of U.S. workers over about ten years; in October 2026 Fortune reported Goldman’s updated estimate of up to about 9%. Axis Intelligence Research calculates that implies roughly 954,000 to 1.4 million displaced workers a year, against about 160,000 AI-cited cuts a year at the 2026 pace. Goldman expects those losses to bunch up in the next downturn.
What is the AI Displacement Visibility Ratio, and how should I cite it?
The AI Displacement Visibility Ratio (ADVR) is the share of all U.S. layoffs and discharges that employers publicly attribute to AI: Challenger AI-cited cuts divided by BLS JOLTS layoffs and discharges. It was 0.26% in 2025 and 0.84% in January–August 2026. Cite as: Axis Intelligence Research, AI Job Displacement Statistics, 2026 (CC BY 4.0).
Correction, October 29, 2026: An earlier version of this article, published June 18, 2026, contained errors. It stated that EU AI Act employment provisions took effect August 2, 2026; under the European Commission’s current timeline, high-risk rules for employment apply from December 2, 2027. It stated Goldman Sachs projected an unemployment effect of no more than 0.5 percentage points; Goldman’s March 2026 analysis gives 0.6 points for a decade-long transition. It described Goldman’s 16,000-a-month estimate as a measured count built on an employer survey of 1,200 firms; Fortune’s report of the note describes a regression-based estimate, and no such survey is described. Goldman’s displacement estimate has since been raised to up to about 9% of workers (Fortune, October 22, 2026). It attributed a “nearly 20%” decline among 22- to 25-year-old developers to the Stanford AI Index 2026, which reports no such figure; the related Stanford Digital Economy Lab study now reports a 19% gap for 22- to 25-year-olds in AI-exposed occupations. It described the August 2025 Stanford study as using payroll data from one software firm; it used ADP payroll data covering millions of workers. Occupation-level “Goldman exposure scores” and automation-risk percentages could not be traced to any Goldman Sachs publication and have been removed, as has the AI Displacement Reality Index (ADRI), whose 2024 and 2026 readings could not be reproduced from published inputs. Report errors to [email protected].