Fifty four percent of the layoff events tracked in 2026 named artificial intelligence, automation or machine learning as a contributing factor, against fewer than 8 percent of announcements in 2025.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source Over the same period, Challenger, Gray and Christmas counted 477,033 announced US job cuts through July, down 41 percent from the 806,383 announced in the first seven months of 2025, with 112,713 of this year's cuts attributed to AI.2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source The stake is that the figure now shaping political and investor readings of AI displacement measures what employers say about their cuts, not how many people lost work. We assess with moderate confidence that the AI labor effect actually visible in payroll data this year runs through suppressed hiring of young workers rather than through the layoffs that carry its name, and that the attribution share and the job loss level are diverging because the two series answer different questions.
The 54 percent and the 24 percent are not the same statistic
The headline share is event weighted. It counts announcements, so a company cutting 40 people and a company cutting 16,000 each contribute one observation, and the resulting 54 percent describes how often AI appears in the reasoning, not how much of the workforce reduction it explains.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source Challenger's number is worker weighted. Of 477,033 announced cuts through July, 112,713 were attributed to AI, roughly 24 percent, and in July alone AI accounted for 10,970 of 33,429 cuts, about a third of the month.2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source Both numbers can be correct at once. A reader who moves between them without noticing the denominator will overstate displacement by a factor of two.
The trackers also disagree with each other on the level. IBTimes, drawing on Layoffs.fyi, puts AI attributed cuts above 170,000 workers year to date.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source A separate tracker compiled by ResumePulse puts the figure at 205,000 through August and notes that this matches the full year 2025 total in under eight months.5 Outsource Accelerator 2026-08-25 Citing the ResumePulse August 2026 tracker: AI attributed layoffs in the United States reached 205,000 workers through August 2026, matching the full year 2025 figure in under eight months; cuts concentrated in customer service, data operations, entry level software roles and finance back offices; the share of organisations naming AI as their stated layoff reason has doubled since January 2026. Open source Challenger's own count is 112,713.2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source None of these is a government series. All three are private compilations of press releases, WARN filings and news coverage, with different inclusion rules, and the spread between the highest and lowest is roughly the size of the entire technology sector's announced cuts this year.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source 2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source 5 Outsource Accelerator 2026-08-25 Citing the ResumePulse August 2026 tracker: AI attributed layoffs in the United States reached 205,000 workers through August 2026, matching the full year 2025 figure in under eight months; cuts concentrated in customer service, data operations, entry level software roles and finance back offices; the share of organisations naming AI as their stated layoff reason has doubled since January 2026. Open source That is a caveat about the evidence, not a footnote to it.
The line that is falling
The aggregate direction is the part that gets lost. Announced cuts in July 2026 were 33,429, down 27 percent from June, while announced hiring plans reached 16,095 for the month and 107,500 year to date, up 25 percent on 2025.2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source Technology is the exception that proves the concentration: 149,023 cuts year to date, up 67 percent on last year, inside a national total that fell by two fifths.2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source Andy Challenger's reading is that AI is "shifting the labor market, it is not dismantling it".2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source
Indeed's posting data describes the same shape from the demand side. The Job Postings Index sat at 101.8 in mid August, marginally above its February 2020 baseline, with new postings at 97.2 and total postings down 2.9 percent year over year.4 Indeed Hiring Lab 2026-08-24 Job Postings Index at 101.8 with new postings at 97.2 and total postings down 2.9 percent year over year; software development postings at 74.4 after a low of 61.1 in May 2025; AI related postings at 6.3 percent of all postings against a 3.3 percent peak in 2022; hiring rate 3.4 percent, quits 2 percent, layoffs 1.1 percent; posted wages up 2.5 percent against 3.4 percent inflation; real private sector wages down 0.4 percent year over year in Q2 2026, the first decline since 2022. Open source Software development postings, the most AI exposed white collar category with a clean series, stood at 74.4, well below pre pandemic levels though recovered from a low of 61.1 in May 2025.4 Indeed Hiring Lab 2026-08-24 Job Postings Index at 101.8 with new postings at 97.2 and total postings down 2.9 percent year over year; software development postings at 74.4 after a low of 61.1 in May 2025; AI related postings at 6.3 percent of all postings against a 3.3 percent peak in 2022; hiring rate 3.4 percent, quits 2 percent, layoffs 1.1 percent; posted wages up 2.5 percent against 3.4 percent inflation; real private sector wages down 0.4 percent year over year in Q2 2026, the first decline since 2022. Open source AI related postings reached 6.3 percent of all postings, roughly double their 2022 peak of 3.3 percent.4 Indeed Hiring Lab 2026-08-24 Job Postings Index at 101.8 with new postings at 97.2 and total postings down 2.9 percent year over year; software development postings at 74.4 after a low of 61.1 in May 2025; AI related postings at 6.3 percent of all postings against a 3.3 percent peak in 2022; hiring rate 3.4 percent, quits 2 percent, layoffs 1.1 percent; posted wages up 2.5 percent against 3.4 percent inflation; real private sector wages down 0.4 percent year over year in Q2 2026, the first decline since 2022. Open source The hiring rate was 3.4 percent, the quits rate 2 percent and the layoffs rate 1.1 percent.4 Indeed Hiring Lab 2026-08-24 Job Postings Index at 101.8 with new postings at 97.2 and total postings down 2.9 percent year over year; software development postings at 74.4 after a low of 61.1 in May 2025; AI related postings at 6.3 percent of all postings against a 3.3 percent peak in 2022; hiring rate 3.4 percent, quits 2 percent, layoffs 1.1 percent; posted wages up 2.5 percent against 3.4 percent inflation; real private sector wages down 0.4 percent year over year in Q2 2026, the first decline since 2022. Open source That combination is a market where few people are being fired and few are being hired, which is precisely the environment in which a modest number of AI attributed cuts can dominate the narrative.
Where the displacement is actually measurable
The strongest evidence on this question is not a layoff tracker. Using ADP payroll records through June 2026, Erik Brynjolfsson, Bharat Chandar and Ruyu Chen find no evidence of widespread, economy wide displacement, but report that employment among workers aged 22 to 25 in highly AI exposed occupations now sits about 19 percent below where it would be had it tracked similarly aged workers in less exposed roles, a gap that widened from 15 percent in July 2025.3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source The mechanism they identify is the important part: the adjustment runs mainly through reduced hiring rather than increased separations, and it splits along codified versus tacit knowledge, with employment flat or rising for experienced workers in occupations where AI complements rather than substitutes.3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source Pay has not yet moved; the adjustment shows up in employment rather than base wages.3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source
We assess with moderate confidence that these two bodies of evidence are consistent and are being read as if they were rival claims. A hiring freeze aimed at a 22 year old generates no layoff announcement, no WARN filing and no tracker entry, so it is invisible in the 54 percent while being the dominant effect in the payroll data.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source 3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source The publicised statistic and the consequential one are close to disjoint.
Who gains and who loses
Executives gain first. Attributing a reduction to AI reframes a cost decision as a capability decision, and the reframing is cheap: Cisco cut 4,000 roles in May in a quarter it reported 15.8 billion dollars of revenue.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source AI vendors gain second, because every attributed layoff is unpaid evidence of deployment at scale, evidence no customer reference programme could buy. Experienced workers in tacit knowledge occupations gain third, and measurably: their employment is flat or rising in the same series that shows the young cohort falling behind.3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source
The losers are specific. The 22 to 25 cohort in exposed occupations carries almost the entire measured effect, and carries it as a job that was never posted rather than one that was lost.3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source Workers already in place lose on price: real wages and salaries for private sector workers fell 0.4 percent year over year in the second quarter of 2026, the first decline since 2022, with posted wages up 2.5 percent against 3.4 percent inflation.4 Indeed Hiring Lab 2026-08-24 Job Postings Index at 101.8 with new postings at 97.2 and total postings down 2.9 percent year over year; software development postings at 74.4 after a low of 61.1 in May 2025; AI related postings at 6.3 percent of all postings against a 3.3 percent peak in 2022; hiring rate 3.4 percent, quits 2 percent, layoffs 1.1 percent; posted wages up 2.5 percent against 3.4 percent inflation; real private sector wages down 0.4 percent year over year in Q2 2026, the first decline since 2022. Open source Anyone laid off for ordinary reasons loses twice, because a 3.4 percent hiring rate makes re entry slow regardless of why the cut happened.4 Indeed Hiring Lab 2026-08-24 Job Postings Index at 101.8 with new postings at 97.2 and total postings down 2.9 percent year over year; software development postings at 74.4 after a low of 61.1 in May 2025; AI related postings at 6.3 percent of all postings against a 3.3 percent peak in 2022; hiring rate 3.4 percent, quits 2 percent, layoffs 1.1 percent; posted wages up 2.5 percent against 3.4 percent inflation; real private sector wages down 0.4 percent year over year in Q2 2026, the first decline since 2022. Open source And the public statistical record loses, since the most cited number on the subject is assembled from press releases by firms with a commercial interest in the tracking.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source 5 Outsource Accelerator 2026-08-25 Citing the ResumePulse August 2026 tracker: AI attributed layoffs in the United States reached 205,000 workers through August 2026, matching the full year 2025 figure in under eight months; cuts concentrated in customer service, data operations, entry level software roles and finance back offices; the share of organisations naming AI as their stated layoff reason has doubled since January 2026. Open source
The counter-case
The strongest argument against this reading is that attribution is honest and simply lagging. Employers have every reason to under report AI as a cause, not over report it, because saying so invites litigation risk, union grievance and political attention, so a jump from under 8 percent to 54 percent might reflect firms finally admitting what was already true.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source On that reading, the falling aggregate is a business cycle artefact and the rising share is the real signal.
The incentive runs the other way too, and one of the industry's own principals says so. Sam Altman has warned of "some AI washing where people are blaming AI for layoffs", and Deutsche Bank analysts have described the same relabelling of ordinary redundancies.1 IBTimes UK 2026-08-26 54 percent of 2026 layoff events cited AI, automation or machine learning against fewer than 8 percent of 2025 announcements; more than 170,000 workers affected by AI attributed cuts on Layoffs.fyi data; more than 165,000 tech roles eliminated in seven months; Cisco cut 4,000 roles in May 2026 in a quarter with 15.8 billion dollars of revenue; Sam Altman on AI washing in layoff attribution; Deutsche Bank on redundancies relabelled as AI. Open source For the attribution to be trustworthy, three things would have to hold: worker weighted AI cuts would have to keep rising as a share of a falling total rather than as a share of events, the young worker employment gap would have to start closing as displacement moved from hiring to separations, and the separations channel in the payroll data would have to open at all.2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source 3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source None of the three is visible today, which is why we hold the assessment at moderate rather than high confidence.
What to watch
- The worker weighted share, not the event share. Challenger's AI attributed cuts stood at about 24 percent of all cuts year to date and 33 percent in July.2 Challenger, Gray and Christmas 2026-08-06 33,429 announced US job cuts in July 2026, down 27 percent from June; 477,033 year to date, down 41 percent from 806,383 in the same period of 2025; technology 149,023 year to date, up 67 percent; AI attributed cuts of 10,970 in July and 112,713 year to date, about 24 percent of all cuts; announced hiring plans of 107,500 year to date, up 25 percent; Andy Challenger on AI shifting rather than dismantling the labor market. Open source If the year to date share clears 35 percent by the December 2026 report while the total keeps falling, attribution is tracking something real; if it stalls near a quarter while the event share climbs past 60 percent, the gap is narrative.
- The separations channel opening. The Stanford series attributes the young worker gap to reduced hiring, not separations.3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source A future update showing elevated separations for the 22 to 25 cohort would be the first hard evidence that AI is removing incumbents rather than closing doors, and would change the policy question entirely.
- The gap crossing into the next cohort. Watch whether the 19 percent shortfall stays confined to ages 22 to 25 or appears among 26 to 30 year olds in the first half of 2027.3 Stanford Digital Economy Lab 2026-08-12 August 2026 update of Canaries in the Coal Mine using ADP payroll records through June 2026: no evidence of widespread economy wide displacement, but employment among workers aged 22 to 25 in highly AI exposed occupations about 19 percent below its counterfactual, widened from 15 percent in July 2025; adjustment operates through reduced hiring rather than separations; declines concentrated where AI substitutes rather than complements; effect in employment rather than base pay. Open source Spread would mean AI is climbing the experience curve, not just blocking its bottom rung.
- Software development postings. The Indeed index for that category was 74.4 in August against a 100 baseline.4 Indeed Hiring Lab 2026-08-24 Job Postings Index at 101.8 with new postings at 97.2 and total postings down 2.9 percent year over year; software development postings at 74.4 after a low of 61.1 in May 2025; AI related postings at 6.3 percent of all postings against a 3.3 percent peak in 2022; hiring rate 3.4 percent, quits 2 percent, layoffs 1.1 percent; posted wages up 2.5 percent against 3.4 percent inflation; real private sector wages down 0.4 percent year over year in Q2 2026, the first decline since 2022. Open source A recovery through 90 during 2027 would falsify the structural reading; a fresh slide toward the May 2025 low of 61.1 would confirm it.
- Whether real wages keep falling. Private sector real wages fell 0.4 percent in the second quarter, the first decline since 2022.4 Indeed Hiring Lab 2026-08-24 Job Postings Index at 101.8 with new postings at 97.2 and total postings down 2.9 percent year over year; software development postings at 74.4 after a low of 61.1 in May 2025; AI related postings at 6.3 percent of all postings against a 3.3 percent peak in 2022; hiring rate 3.4 percent, quits 2 percent, layoffs 1.1 percent; posted wages up 2.5 percent against 3.4 percent inflation; real private sector wages down 0.4 percent year over year in Q2 2026, the first decline since 2022. Open source Two more consecutive negative quarters would move this story out of the hiring channel and into the pay channel, which is where labor politics actually starts.
The number that will get quoted through the autumn is 54 percent. The number that will decide what happens to a generation of entrants is 19 percent, and it is built from payroll records rather than press releases.