AI Workplace Adoption & Job Displacement / Creation

Companion Excel Workbook: AI Workplace Adoption

AI-Attributed Layoffs

AI was the leading cited reason for U.S. announced job cuts for five straight months, March through July 2026, peaking at 40% of May’s cuts; in August and September the share fell below 10%.

U.S. announced job cuts, annual 2023 to 2025 and monthly January to September 2026, with the AI-attributed share rising from 0.6 percent to a 39.8 percent peak in May 2026 and falling to 6.5 percent in August and 9.2 percent in September

U.S. Announced Job Cuts: Total and AI-Attributed Share | Source: Challenger, Gray & Christmas monthly job-cut reports and year-end PDFs (2023 through September 2026).

AI-attributed layoffs rose from 0.6% of total announced cuts in 2023 to about 25% in March and April 2026 and a record 39.8% in May (38,579 cuts), held near a third in June and July, then fell to 6.5% in August (3,462 cuts, the lowest monthly total since December 2025) and 9.2% in September (3,961 cuts), when AI ranked fifth among cited reasons.[1] Through September, AI has been cited for 120,136 cuts, 21.0% of the 2026 total and still the leading reason for the year, more than double the 54,836 attributed to AI in all of 2025.

Overall 2026 layoffs through September (573,195) are down 39% from the same period in 2025, but the comparison is distorted by federal DOGE cuts that defined Q1 2025.[2] Excluding government, layoffs are running about 15% below 2025; September’s 43,281 was the lowest September total since 2022, and third-quarter announcements (129,591) were down 43% from the second quarter.

Technology has accounted for 29% of all 2026 announced cuts through September (165,925 of 573,195), up 54% from the same period of 2025 and more than double the sector's 13% share of the 2025 total; September’s 10,799 technology cuts were up 77% from August’s 6,103, which had been the sector’s lowest month of the year.[3]

Challenger's "AI" category aggregates three distinct causes: actual AI role replacement, AI capex spending crowding out personnel budgets, and "AI washing" of layoffs done for other reasons. The drop in AI attributions since May may also reflect two forces outside the layoffs themselves: public sentiment toward AI has turned more negative in the current political moment, and "AI washing" has drawn pushback, so companies have less reason to name AI in an announcement. See Detailed context and analysis below.

▶ Detailed context and analysis: AI-attributed layoffs and what they actually measure

The "Artificial Intelligence" reason code, first introduced to the Challenger, Gray & Christmas ("Challenger") layoff estimates in May 2023, grew from a rounding error in its first year to the single largest cited reason for layoffs in every month from March through July 2026, before dropping to fourth place in August and fifth in September, behind market conditions, closings, demand downturn and restructuring. Technology companies accounted for 40% of all layoffs in April 2026 (Meta, Microsoft and Block among the announcements) and 39% in May, the sector’s heaviest month since August 2024.[4] The headline AI layoff figure aggregates three distinct dynamics that Challenger does not separately quantify: actual AI role replacement, AI capex spending crowding out personnel budgets, and AI washing.

The technology concentration. Part of the rise in the AI share of total layoffs reflects a compositional shift: the technology sector is making up an unusually large share of all 2026 layoffs, and technology is also where AI is most directly substituting for human work. Challenger reported 165,925 technology cuts in the first nine months of 2026, up 54% from 107,878 in the same period of 2025.[3] As a share of total year-to-date announced cuts, technology rose from 11.4% in 2025 to 28.9% in 2026, a 2.5x increase.[3] Strip out federal-government layoffs (which dominated Q1 2025 due to DOGE actions) and the comparison moderates: technology made up 16.7% of all non-government announced cuts through September 2025, versus 30.2% through September 2026, a 1.8x increase rather than 2.5x.[5] BLS's own monthly data shows a parallel pattern in realized employment: the Information supersector, which includes software publishing, telecommunications, and data processing services, is down 376,000 jobs (12.1%) from its November 2022 peak, while the broader U.S. unemployment rate has stayed between 3.4% and 4.5% over the same period and stood at 4.2% in September 2026.[6] Information employment fell another 10,000 in September 2026, in a month when total nonfarm employment rose only 29,000 and July was revised to a loss.

Actual AI role replacement. The clearest case of AI directly displacing work is in software engineering, where code-generation tools have meaningfully shortened the time required for routine development tasks. Andy Challenger, the firm's chief revenue officer, has identified this as where the role-replacement effect is most concentrated, with large technology firms the primary employers affected. In the one month Challenger broke out AI cuts by industry (September 2025), the technology sector accounted for the entire AI-attributed total. Other knowledge-work categories such as customer support, content production, paralegal work, and basic financial analysis have seen some early displacement but on a much smaller scale.

AI capex spending crowding out personnel budgets. The largest U.S. technology companies are spending heavily on AI infrastructure, with announced 2026 hyperscaler capex running into the hundreds of billions. Andy Challenger put it plainly in May 2026: "Regardless of whether individual jobs are being replaced by AI, the money for those roles is."[7] In this framing, the layoff is real and the AI attribution is also real, but the mechanism is budgetary substitution rather than task automation. A company funding a multi-billion-dollar data-center buildout must find the budget somewhere, and headcount is typically the most flexible line item. This dynamic shows up in the AI-cited reason code regardless of whether any specific eliminated role's tasks were actually being automated.

AI washing. The third category captures cases where companies frame restructurings for other reasons as AI-driven, because the narrative is rewarded by investors and analysts. Challenger himself flagged the January 2026 Amazon layoffs as a likely example: CEO Andy Jassy attributed the 16,000 corporate job cuts to AI, but Challenger thought they looked "more due to over-hiring and reducing layers than to the new technology."[8] Distinguishing AI washing from genuine AI-driven cuts is difficult from the outside, because the same announcement can be all three things at once: a company that over-hired during the post-pandemic boom, faces margin pressure, is redirecting budget to AI capex, and believes some of the eliminated roles will soon be automated. Challenger added a "Technological Update (possibly AI)" reason code in July 2026 for announcements that cite new technology and allude to AI without tying the cuts to it, and noted that the incentive to name AI is shifting: "Naming AI in a layoff announcement can win over investors while pushing current and prospective employees away," and "as regulations start to take shape, companies will be even more careful in their announcements, which would make tracking the impact of AI on jobs more opaque."[12] The August drop in AI attributions should be read with that in mind.

Quantifying the Effect of AI on Jobs: Early Returns

Since January 2022, total U.S. employment has grown 6.0% while the three pure knowledge-work sectors have grown 0.6% combined, and Information employment has declined outright.

Seven-line chart of employment indexed to January 2022 through September 2026: total nonfarm up 6.0 percent, Professional and Business Services and Financial Activities up 1 to 3 percent, Information down 8.2 percent on BLS data and 3.8 percent on ADP data

Knowledge-Sector Employment vs Total Nonfarm: % Change Since January 2022 (through September 2026) | Sources: U.S. Bureau of Labor Statistics, Current Employment Statistics, via FRED; ADP National Employment Report history file. Both seasonally adjusted. Solid lines BLS, dashed lines ADP.

Bureau of Labor Statistics data show that from January 2022 through September 2026, total nonfarm employment is up 6.0% (+9.0 million jobs), while the three pure knowledge-work job categories rose 0.6% combined.[9] Information fell 8.2% (−245,000); Financial Activities rose 1.7% (+148,000); Professional and Business Services rose 1.3% (+299,000).

ADP's private-payroll series, compiled from a separate dataset and method, shows the same pattern over the same period: Information down 3.8%, Financial Activities up 3.2%, Professional and Business Services up 1.8%.[10] The two sources differ on magnitude but agree on direction for all three sectors.

Through the first nine months of 2026, BLS Information employment fell by 103,000 jobs year to date, continuing the decline that began after its November 2022 peak; the September data released October 2 showed a further 10,000 drop and revised July’s total nonfarm change to a loss. ADP's September report (September 30) shows Information up 26,000 year to date, roughly flat, while Professional and Business Services (−11,000) and Financial Activities (−16,000) were the only ADP sectors to shed jobs in the month.

These payroll trends are consistent with the announced-layoff data in the section above. Challenger job-cut announcements and their AI-attributed share rose through 2025 and the first half of 2026 as knowledge-sector payrolls stagnated or declined. The August and September drop in AI attributions has no counterpart yet in the payroll data, where Information has fallen in seven of the nine months of 2026.

We are still in the early stages of the AI labor market transformation, and initial predictions regarding vulnerable roles have proven mostly accurate so far. See following section for more details on which jobs and job categories are most susceptible to AI replacement.

Early-Career Workers in AI-Exposed Occupations

In ADP payroll data, employment of 22-to-25-year-olds in the most AI-exposed occupations is 13% below its November 2022 level, while their peers in the least-exposed occupations are up 8% and older workers in the same exposed occupations are up 9% to 5%.

Four-line chart of an employment index set to 100 in November 2022: workers aged 22 to 25 in the most AI-exposed occupation quintile fall to 87 by August 2026, while 22 to 25 year olds in the least-exposed quintile rise to 108 and workers aged 41 to 49 and 50 plus in the most-exposed quintile rise to 109 and 105

Employment Index by Age and AI Exposure, Nov 2022 = 100 (through August 2026) | Source: Stanford Digital Economy Lab, AI Economic Indicators, Canaries Dashboard with ADP Research. Solid lines ages 22 to 25; dashed lines older workers in the most-exposed quintile.

The "Canaries in the Coal Mine" series is the most direct public test of whether AI is changing hiring. Stanford's Digital Economy Lab (Brynjolfsson, Chandar and Chen) tracks payroll employment at a fixed sample of ADP client firms by age group and by occupational AI exposure. Since ChatGPT's release, employment of 22-to-25-year-olds in the most-exposed fifth of occupations has fallen to 87.2 (Nov 2022 = 100) while the least-exposed fifth has risen to 108.1; workers aged 41 to 49 in the same most-exposed occupations stand at 109.4.[13]

The August 2026 revision of the paper (ADP data through June 2026) puts early-career employment in AI-exposed jobs 19% below where it would be had it kept pace with less-exposed peers, up from 13% a year earlier, and finds the adjustment is coming through reduced hiring rather than layoffs, concentrated where AI substitutes for tasks rather than complements them. The authors describe these as early, descriptive indicators, not causal estimates, and the sample is ADP client firms rather than the whole economy.[13]

For CRE the hiring channel matters more than the layoff channel: entry-level knowledge workers are the tenants of starter apartments and the occupants of the desks that fill out a floor plate. A gap that widens through hiring freezes will show up in office and multifamily demand more slowly, and less visibly, than an announced layoff. The dashboard updates monthly and is the series CRE42 will follow into 2027.

AI Exposure by Occupation

Automation exposure is highest in office support, business and finance, and computer and math, while the largest occupation families by headcount sit in the lower-exposure service and production groups.

U.S. employment versus AI automation exposure by occupation family, sorted by exposure

U.S. Employment vs. AI Automation Exposure, by Occupation Family | Source: BLS OEWS (May 2025) for employment; automation scores from the Harvard Business School / OpenAI occupation-vulnerability model, mapped by CRE42.

The analysis scores 772 occupation units, covering roughly 153 million jobs (about 98% of U.S. employment) across 13 occupation families, for exposure to generative AI. Of these, 228 units and 49.6 million jobs fall in the seven knowledge-work families.[11]

Exposure scores come from a Harvard Business School study (Srinivasan, Chen, and Zakerinia, 2024; updated 2025) that used OpenAI's GPT-4o model to rate each occupation's automation and augmentation potential from its underlying O*NET task content. CRE42 mapped those scores to BLS OEWS May 2025 employment at the six-digit SOC level.

Each occupation is assigned to one of four exposure bands by automation score: Low (below 0.20), Moderate (0.20 to 0.35), Elevated (0.35 to 0.50), and High (0.50 and above). Knowledge work concentrates at the top: 55% of knowledge-work employment sits in the High band, versus 21% of total U.S. employment.

Among the 16 high-exposure occupations with more than 500,000 jobs each, combined employment fell 240,730 from May 2024 to May 2025 (table below); the same group was roughly flat over the longer 2022 to 2025 window. Whether 2025 marks an inflection or a noisy year is unresolved; CRE42 will update this figure as each annual OEWS release lands.

High-exposure occupations (automation score ≥ 0.50) with more than 500,000 jobs, ranked by automation score. The 16 occupations shown lost a combined 240,730 jobs from May 2024 to May 2025, even though they grew on net over the longer 2022–2025 window.
Occupation	Job family	Automation score	2025 jobs	Change 2022–2025		Change 2024–2025		Band
Occupation	Job family	Automation score	2025 jobs	Jobs	%	Jobs	%	Band
Project Management Specialists	Business and Finance	0.70	1,066,670	+222,760	+26.4%	+60,510	+6.0%	High
Bookkeeping, Accounting, and Auditing Clerks	Office Support	0.68	1,373,680	−177,070	−11.4%	−82,090	−5.6%	High
Medical Secretaries and Administrative Assistants	Office Support	0.66	961,610	+278,980	+40.9%	+130,850	+15.8%	High
Receptionists and Information Clerks	Office Support	0.64	910,180	−100,990	−10.0%	−54,350	−5.6%	High
Secretaries and Administrative Assistants, Except Legal, Medical, and Executive	Office Support	0.62	1,706,790	−119,920	−6.6%	−31,030	−1.8%	High
Office Clerks, General	Office Support	0.60	2,464,940	−52,410	−2.1%	−45,610	−1.8%	High
Shipping, Receiving, and Inventory Clerks	Office Support	0.59	816,870	−31,370	−3.7%	−40,760	−4.8%	High
Business Operations Specialists, All Other	Business and Finance	0.58	1,087,090	+5,860	+0.5%	−41,110	−3.6%	High
Human Resources Specialists	Business and Finance	0.58	912,430	+77,070	+9.2%	−5,030	−0.5%	High
Customer Service Representatives	Office Support	0.57	2,595,750	−284,090	−9.9%	−130,180	−4.8%	High
First-Line Supervisors of Office and Administrative Support Workers	Office Support	0.55	1,436,680	−58,760	−3.9%	−58,900	−3.9%	High
Computer Systems Analysts	Computer and Math	0.55	519,530	+14,320	+2.8%	+21,730	+4.4%	High
Market Research Analysts and Marketing Specialists	Business and Finance	0.54	899,580	+100,960	+12.6%	+38,440	+4.5%	High
Managers, All Other	Management	0.53	622,190	+78,900	+14.5%	−8,790	−1.4%	High
Accountants and Auditors	Business and Finance	0.51	1,449,500	+47,080	+3.4%	+1,210	+0.1%	High
Management Analysts	Business and Finance	0.50	898,280	+89,420	+11.1%	+4,380	+0.5%	High
Total (16 occupations)			19,721,770	+90,740	+0.5%	−240,730	−1.2%

Explore all 771 occupations interactively in the AI Exposure Explorer →

Of the 20 highest-scoring occupations, 16 saw employment decline between 2022 and 2025, and 15 declined in the single year from 2024 to 2025.

Footnotes

[1] Challenger, Gray & Christmas, monthly job-cut reports (2023–2026): 2023 Annual via Sep 2024 PDF, 2024 Annual via Sep 2024 "AI Cuts by Industry" table cross-confirmed via 2025 cumulative arithmetic, 2025 Year-End Report (Jan 8 2026), and monthly releases Feb 5, Mar 5, Apr 2, May 7, Jun 4, Jul 1, Aug 6, Sep 3 and Oct 1 2026 (Table 4, Job Cuts by Reason, in each monthly PDF; the May release calls 38,579 "the highest monthly total ever recorded for the reason"). AI was first cited as a reason category in May 2023. challengergray.com (monthly job-cut reports)↩

[2] Challenger, Gray & Christmas, September 2026 release (Oct 1 2026): "Through September, employers have announced 573,195 job cuts, down 39% from the 946,426 cuts announced in the first nine months of 2025. Excluding the Government sector, job cut announcements are down 15% (550,185 versus 646,671 through September 2025)." The same release: September's 43,281 "is the lowest total for the month since 2022," and "Third quarter layoff plans totaled 129,591, down 43% from the 226,242 layoff plans in Q2 2026." Q1 2025 figures were inflated by roughly 282,000 federal government cuts attributed to Department of Government Efficiency actions. challengergray.com (Sep 2026 PDF)↩

[3] Challenger, Gray & Christmas, September 2026 PDF (Oct 1 2026): "Technology companies announced 10,799 cuts in September, up 77% from 6,103 in August," for a year-to-date total of 165,925, up 54% from 107,878 through September 2025 (Table 2); "Technology accounts for 29% of all job cuts announced in 2026." 2025 full-year Technology total of 154,445 from Challenger 2025 Year-End Report (Jan 8 2026). Shares computed against the 2025 grand total of 1,206,374 (12.8%) and the 2026 year-to-date grand total of 573,195 (28.9%). challengergray.com (Sep 2026 PDF)↩

[4] Computed from Challenger monthly PDFs: April 2026, 33,361 Technology ÷ 83,387 Total = 40.0%; May 2026, 38,242 ÷ 97,006 = 39.4%. The May release: "Technology announced 38,242 job cuts in May, the highest monthly total for the sector since August 2024, when 39,563 cuts were recorded."↩

[5] Computed from Challenger September 2026 PDF Table 2 (job cuts by industry, year-to-date columns). 2025 through September: 946,426 total cuts − 299,755 government = 646,671 ex-government; tech 107,878 ÷ 646,671 = 16.7%. 2026 through September: 573,195 − 23,010 = 550,185 ex-government; tech 165,925 ÷ 550,185 = 30.2%. Multiplier 30.2% ÷ 16.7% = 1.81x. The same calculation on the January to August windows (Aug 2026 PDF) gave 17.1% and 30.6%, 1.78x, and on January to April (Apr 2026 PDF) 20.0% and 29.5%, 1.47x. challengergray.com (Sep 2026 PDF)↩

[6] U.S. Bureau of Labor Statistics, Current Employment Statistics and Current Population Survey, seasonally adjusted, September 2026 vintage (Employment Situation released Oct 2 2026, Table B-1 and Table A-1; history via FRED series USINFO, PAYEMS and UNRATE). Information: 3,115,000 in November 2022, 2,739,000 in September 2026 (−376,000, −12.1%); August to September 2026 −10,000. Total nonfarm August to September 2026 +29,000, with July revised from +21,000 to −10,000 and August from +162,000 to +133,000. Unemployment rate: low 3.4% (April 2023), high 4.5% (November 2025), 4.2% in September 2026. bls.gov ↩

[7] Andy Challenger, quoted in Challenger, Gray & Christmas April 2026 release (May 7 2026). challengergray.com ↩

[8] Challenger, Gray & Christmas commentary on Amazon's January 2026 layoff announcement; Amazon disclosed approximately 16,000 corporate job cuts in late January 2026 with CEO Andy Jassy citing AI as a driver. challengergray.com (Jan 2026 PDF)↩

[9] U.S. Bureau of Labor Statistics, Current Employment Statistics (Establishment Survey), seasonally adjusted, January 2022 through September 2026, via FRED series PAYEMS (total nonfarm), USINFO (Information), USFIRE (Financial Activities), and USPBS (Professional and Business Services). July, August and September 2026 figures from the Employment Situation released October 2, 2026 (Table B-1); earlier months as revised in the September 4 vintage. bls.gov/ces ↩

[10] ADP Research / Stanford Digital Economy Lab, ADP National Employment Report, private-sector employment, seasonally adjusted, Industry series (Information; Financial activities; Professional and business services) from the ADP NER history file (October 2026 vintage), which carries the September 2026 report released September 30, 2026. ADP covers private employers only and is built from anonymized payroll records of more than 25 million U.S. employees. adpemploymentreport.com ↩

[11] CRE42 groups the 13 SOC major occupation families into three work types. Knowledge work (7 families): Management; Business and Finance; Computer and Math; Engineering; Life, Physical, and Social Science; Legal; and Office Support. Hybrid (4 families): Arts, Sports, and Media; Educational Instruction; Health Care; and Sales. Physical (2 families): Services and Other; and Production, Construction, and Transportation.↩

[12] Challenger, Gray & Christmas, July 2026 release (Aug 6 2026): "Challenger uses this categorization ["Technological Update (possibly AI)"] when a company states new technology was the main reason for layoffs and AI is alluded to but not directly tied to the cuts. Challenger tracked 20,219 cuts for this reason in 2025." Both quotations are Andy Challenger in the same release. challengergray.com (Jul 2026 release)↩

[13] Stanford Digital Economy Lab, AI Economic Indicators, Canaries Dashboard (with ADP Research): canaries_age_by_exposure.csv, vintage 2026-09-17, data through August 2026; employment index normalized to 100 at November 2022 for each age-by-exposure series, exposure quintiles per Eloundou, Manning, Mishkin and Rock (2024). Paper: Brynjolfsson, Chandar and Chen, "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence," revised August 12, 2026 (ADP data through June 2026): "employment for early-career workers (ages 22-25) in the most AI-exposed occupations now stands 19% below" the kept-pace counterfactual; most-exposed two quintiles down about 11% and least-exposed three quintiles up about 10% from November 2022 to June 2026. digitaleconomy.stanford.edu (Canaries Dashboard); paper (PDF, Aug 2026)↩

Companion workbook. technology-knowledge-ai-workplace-adoption.xlsx — Challenger job-cut data with AI attribution, BLS/ADP monthly knowledge-sector employment, HBS/OpenAI occupation automation scores mapped to BLS OEWS, and 2022–2025 employment-change analysis, and the Stanford/ADP Canaries early-career employment index