# AI Job Displacement Register — isaigonnatakemyjob.com

Updated 2026-08-18. Licensed CC BY 4.0. Canonical page: https://isaigonnatakemyjob.com/

A register of layoffs where the employer named AI, occupations with measured decline, and companies AI disruption killed. Every entry carries a source URL. Nothing here is our own estimate; projections are labelled and attributed.

## The short answer

As of August 2026, AI has not taken most jobs. It has measurably taken specific tasks inside many jobs, the entry rung of several white-collar careers, and the business models of companies whose product was information. Employers attributed 112,713 of 477,033 US job cuts announced January–July 2026 to AI (Challenger, Gray & Christmas). Employment for 22–25 year-olds in the most AI-exposed occupations is about 19% below its counterfactual (Stanford Digital Economy Lab). Economy-wide, the Budget Lab at Yale finds no substantial labour-market disruption yet.

Scope: 45 occupations · 50 layoff announcements · 37 companies.

## Headline statistics

- **1,206,374** — Total US announced job cuts, 2025 (vs 761,358 in 2024 (+58%); highest annual total since 2020 Q4 2025 was the highest Q4 for layoffs since 2008; YTD hiring plans the lowest since 2010.)
  Source: Challenger, Gray & Christmas https://www.challengergray.com/blog/2025-year-end-challenger-report-highest-q4-layoffs-since-2008-lowest-ytd-hiring-since-2010/
- **54,836** — US job cuts explicitly attributed to AI, 2025 (5th-largest stated reason in 2025, behind DOGE actions (293,753), market/economic conditions (253,206), closings (191,480) and restructuring (133,611) October 2025 was the peak month at 31,039 AI-attributed cuts. A separate 'Technological Update' category (which includes some AI implementation) accounted for ~20,000 additional cuts through May 2025.)
  Source: Challenger, Gray & Christmas https://www.challengergray.com/blog/2025-year-end-challenger-report-highest-q4-layoffs-since-2008-lowest-ytd-hiring-since-2010/
- **112,713** — US job cuts explicitly attributed to AI, Jan-Jul 2026 (AI is the #1 stated reason YTD 2026, ahead of market/economic conditions (90,075), closings (84,630), restructuring (57,476), contract loss (40,758) and cost-cutting (34,702) July 2026 alone: 10,970 AI-attributed cuts out of 33,429 total. Andy Challenger: "Layoff plans continue to be announced primarily in Tech, and artificial intelligence is still the story, as investments in the technology reshape organizations." He added: "Hiring has also increased over last year by 25%, so while AI is shifting the labor market, it is not dismantling it," and "Naming AI in a layoff announcement can win over investors while pushing current and prospective employees away.")
  Source: Challenger, Gray & Christmas https://www.challengergray.com/wp-content/uploads/2026/08/Challenger-Report-July-2026.pdf
- **477,033** — Total US announced job cuts, Jan-Jul 2026 (July 2026 total of 33,429 was the lowest monthly figure in two years AI-attributed cuts = 23.6% of all announced US job cuts YTD 2026.)
  Source: Challenger, Gray & Christmas https://www.challengergray.com/wp-content/uploads/2026/08/Challenger-Report-July-2026.pdf
- **38,579 AI-attributed cuts in May 2026 = 40% of all cuts that month** — Peak month for AI-attributed share of layoffs (YTD through May 2026: 87,714 AI-attributed of 397,755 total (22%) Andy Challenger: "AI is now the leading reason companies give for cutting jobs and the primary industry citing it is Technology.")
  Source: Challenger, Gray & Christmas https://www.challengergray.com/wp-content/uploads/2026/06/Challenger-Report-May-2026.pdf
- **2023 (May-Dec): 3,900 | 2024: 13,089 | 2025: 54,836 | 2026 YTD (Jul): 112,713** — AI-attributed US job cuts by year since tracking began (2025 was ~4.2x 2024; 2024 was ~3.4x the 2023 partial-year figure MEDIUM CONFIDENCE on the 2023 and 2024 annual figures — sourced from a compilation of Challenger's monthly reports rather than a single Challenger publication. Challenger first broke out AI as a separate layoff cause in May 2023 (3,900 cuts, all in tech). Straight News independently reported 'more than 71,000 announced cuts linked to AI since 2023' as of Dec 2025 (https://san.com/cc/artificial-intelligence-tied-to-more-than-50000-layoffs-in-2025/), consistent with 3,900 + 13,089 + 54,836 = 71,825.)
  Source: Challenger, Gray & Christmas (2023/2024 figures via third-party compilation of Challenger reports) https://gist.github.com/mcphil/08f8f011f27c9864d157ead0b9d716d9
- **2023: just under 265,000 | 2024: just over 152,000 across 551 companies | 2025: ~123,000 across 257 companies** — Total tech-industry layoffs by year (global tracker) (2025 was ~20% below 2024 and the lowest total since 2022; 2023 remains the record year Note the divergence: overall tech layoff VOLUME fell in 2025 even as the AI-ATTRIBUTED share rose sharply. 2023's peak was driven by post-COVID overhiring correction, not AI.)
  Source: Layoffs.fyi (reported by Salesforce Ben) https://www.salesforceben.com/how-bad-were-tech-layoffs-in-2025-and-what-can-we-expect-next-year/
- **No substantial disruption detected** — Macro evidence of AI labour-market disruption (US) (Occupational mix shifts began in 2021, predating widespread genAI adoption; unemployed workers across all durations remained in occupations with roughly 25-35% of tasks potentially automatable Finding: "the picture of AI's impact... largely reflects stability, not major disruption" and "measures of exposure, automation, and augmentation show no sign of being related to changes in employment or unemployment." Important counterweight to company-level AI layoff announcements.)
  Source: The Budget Lab at Yale https://budgetlab.yale.edu/research/evaluating-impact-ai-labor-market-novemberdecember-cps-update
- 170 million new jobs created and 92 million displaced by 2030 — a net increase of 78 million jobs, with 22% of all jobs experiencing structural disruption. (Nearly 40% of skills required in jobs will change by 2030. 59 of every 100 workers globally will need reskilling or upskilling, and 11 of those are unlikely to receive it — roughly 120 million workers at medium-term redundancy risk. 41% of employers plan to reduce their workforce as AI automates tasks; 77% plan to upskill.)
  Source: World Economic Forum, Future of Jobs Report 2025 https://www.weforum.org/press/2025/01/future-of-jobs-report-2025-78-million-new-job-opportunities-by-2030-but-urgent-upskilling-needed-to-prepare-workforces/
- Employment for 22–25 year-olds in the most AI-exposed occupations is about 19% below where it would be had it tracked less-exposed peers — a gap that widened from 15% in July 2025. (In absolute terms, employment of 22–25 year-olds in the two most AI-exposed occupational quintiles fell about 11% between November 2022 and June 2026, while the same age group in the three least-exposed quintiles grew about 10%. The authors stress these are descriptive patterns, not causal estimates, and note uncertainty over whether the trend will accelerate, stabilise or reverse.)
  Source: Stanford Digital Economy Lab, 'Canaries in the Coal Mine?' (August 2026 update), Brynjolfsson, Chandar & Chen https://digitaleconomy.stanford.edu/news/canariesaug26/
- Activities accounting for 29.5% of hours worked in the US economy could be automated by 2030 with generative AI — up from 21.5% without it — requiring an additional 12 million occupational transitions. (Declines in food services, customer service and sales, office support, and production work account for almost 10 million (more than 84%) of those 12 million shifts. Office support demand could fall by 1.6 million jobs, retail sales by 830,000, administrative assistants by 710,000 and cashiers by 630,000. Workers in the lowest wage quintiles are up to 14 times more likely to need an occupational change than the highest earners.)
  Source: McKinsey Global Institute, Generative AI and the Future of Work in America https://www.mckinsey.com/mgi/our-research/generative-ai-and-the-future-of-work-in-america
- New-graduate hiring is down roughly 65% at the largest tech companies and about 76% at early-stage startups compared with 2019. (Since ChatGPT's launch, AI/ML engineer share of engineering roles rose 39% and research engineers grew 28%, while front-end engineer roles fell about 25%. SignalFire's reading is that junior roles were cut for cost optimisation while experienced engineers became more valuable through AI leverage — the entry rung was removed, not the ladder.)
  Source: SignalFire, State of Tech Talent Report 2026 https://www.signalfire.com/blog/signalfire-state-of-talent-report-2026
- 30% of US workers could see at least 50% of their occupation's tasks disrupted by generative AI; 85% could see at least 10% affected. (Women are disproportionately exposed: 36% of female workers are in occupations where generative AI could save 50% of task completion time, versus 25% of male workers. Office and administrative support — about 19 million American workers, predominantly women — is among the most exposed groups. Unlike earlier automation waves, this targets non-routine cognitive work in middle- and higher-paid professional roles, while physically intensive blue-collar work faces minimal disruption.)
  Source: Brookings Institution https://www.brookings.edu/articles/generative-ai-the-american-worker-and-the-future-of-work/
- AI was cited in 112,713 of the 477,033 US job cuts announced through July 2026 — roughly 24% of all cuts, and the leading stated reason for five consecutive months. (In July 2026 alone, 10,970 of 33,429 announced cuts (33%) were attributed to AI. Technology was the largest sector at 149,023 cuts (31% of the total), followed by transportation (41,748), health care/products (34,426), services (23,942) and government (20,752). For scale, Challenger tracked only 20,219 cuts under 'Technological Update (possibly AI)' in all of 2025.)
  Source: Challenger, Gray & Christmas, Job Cut Announcement Report (July 2026) https://www.challengergray.com/blog/challenger-report-layoffs-fall-hiring-picks-up-ai-leads-for-fifth-straight-month/
- One in four jobs worldwide (25%) is potentially exposed to generative AI, rising to 34% in high-income countries. (Clerical positions are the most vulnerable occupational group. In high-income countries, 9.6% of female employment falls into the highest automation-risk category versus 3.5% of male employment. The ILO emphasises that 'transformation, not replacement, is the most likely outcome' — job evolution through task modification is far more probable than elimination.)
  Source: International Labour Organization / NASK Global Index https://www.ilo.org/resource/news/one-four-jobs-risk-being-transformed-genai-new-ilo%E2%80%93nask-global-index-shows
- The equivalent of 300 million full-time jobs globally is exposed to AI automation, with roughly two-thirds of US occupations exposed to some degree — but exposure is not loss. (Of exposed roles, a quarter to as much as half of the workload could be replaced, and Goldman estimates a 7% lift to global GDP (about $7 trillion). Goldman notes that technology-driven new occupations accounted for more than 85% of employment growth over the last 80 years. Corroborating the caution: Stanford SIEPR reports that only 5% of firms report any employment impact from AI, split evenly between gains and losses, and that employment grew 10% in the two years following enterprise AI adoption — though recent-graduate unemployment reached 5.6% in early 2026, up 1.6 percentage points from three years prior.)
  Source: Goldman Sachs Research; corroborating data from Stanford SIEPR https://www.goldmansachs.com/insights/articles/generative-ai-could-raise-global-gdp-by-7-percent
- 68.01% of Google searches in the US ended without a click in January-April 2026, up from 60.45% in 2024 - a 7.56 percentage-point rise in two years. Clicks of any kind fell 9.51 percentage points (a 22.9% relative drop) over the same period. In 2016 the figure was around 45%. (zero-click search rate Similarweb desktop and mobile clickstream panel, US)
  Source: SparkToro https://sparktoro.com/blog/in-2026-less-than-one-third-of-google-searches-still-send-a-click/
- When a Google AI Overview appears, users click a traditional search result in only 8% of searches, versus 15% without one. Only 1% of visits result in a click on a link cited inside the AI summary. Users were also more likely to end their browsing session entirely after a page with a summary (26%) than without one (16%). AI Overviews appeared on ~18% of all searches studied. (AI Overview click-through impact Pew Research Center, browsing data from 900 US adults, March 2025)
  Source: Search Engine Land (reporting Pew Research) https://searchengineland.com/google-ai-overviews-hurting-clicks-study-459434
- Small publishers (1,000-10,000 daily page views) lost 60% of their search traffic over two years. Medium publishers (10,000-100,000 daily views) lost 47%; large publishers (100,000+) lost 22%. Google Search page views to publishers fell 34% between December 2024 and December 2025; Google Discover fell 15%. (publisher search traffic decline by publisher size Chartbeat, aggregated from thousands of client websites globally)
  Source: PPC Land (reporting Chartbeat data published by Axios) https://ppc.land/small-publishers-lost-60-of-search-traffic-as-ai-reshapes-the-web/
- Organic search traffic to news sites fell from a peak of over 2.3 billion visits in mid-2024 to under 1.7 billion. Zero-click news searches rose from 56% to nearly 69% between May 2024 (AI Overviews launch) and May 2025. ChatGPT referrals to news sites grew from under 1 million (Jan-May 2024) to more than 25 million in 2025 - a 25x increase that replaced a fraction of the ~600 million lost search visits. (news publisher search referrals vs AI chatbot referrals Similarweb)
  Source: TechCrunch https://techcrunch.com/2025/07/02/chatgpt-referrals-to-news-sites-are-growing-but-not-enough-to-offset-search-declines
- Stack Overflow monthly question volume fell from 207,204 at its March 2014 peak to 1,442 in July 2026 — a 99.3% collapse. The whole of 2025 produced 108,981 questions, fewer than a single month at peak. Meanwhile 84% of developers report using AI tools and 51% of professional developers use them daily. (developer Q&A volume collapse Stack Overflow 2025 Developer Survey (49,000+ respondents, 177 countries) and Stack Exchange question data)
  Source: PPC Land https://ppc.land/stack-overflow-drops-to-1-442-questions-in-july-down-99-from-2014-peak/
- Chegg's revenue fell from $776.3 million in FY2021 (with 7.8 million subscribers, peaking at 8.2 million in 2022) to $376.9 million in FY2025 (-39%), and Q2 2026 revenue of $51.8 million (-51% YoY). Market capitalisation fell from $11.64 billion at the end of 2020 to $167.94 million in May 2026. The company cut 441 jobs in November 2024 and a further 388 (~45% of staff) announced in October 2025. (single-company AI disruption, peak to trough Chegg SEC-filed earnings releases; StockAnalysis.com market cap history)
  Source: Business Wire / Chegg Investor Relations / StockAnalysis.com https://www.businesswire.com/news/home/20260806916599/en/Chegg-Reports-Second-Quarter-2026-Earnings
- Wikipedia human pageviews fell approximately 8% comparing May-August 2025 with the same months in 2024. The Wikimedia Foundation attributed this to 'generative AI and social media' and to 'search engines providing answers directly to searchers, often based on Wikipedia content.' (reference-site traffic decline Wikimedia Foundation pageview analysis)
  Source: Wikimedia Foundation (Diff) https://diff.wikimedia.org/2025/10/17/new-user-trends-on-wikipedia/
- Approximately 140,000 US tech jobs were cut in 2026 through July at companies that explicitly cited AI, including Oracle (21,000, 13%), Amazon (16,000, ~9% of corporate), Dell (11,000, 10%), Meta (8,000, 10%), Microsoft (4,800), Cisco (4,000), PayPal (4,500+, 20%), Block (4,000, ~50%), Intuit (3,000, 17%), Atlassian (1,600, 10%), Cloudflare (1,100, 20%), Snap (1,000, 16%) and Salesforce (under 1,000, citing 'benefits and efficiencies of Agentforce' reducing support needs). (AI-attributed layoffs TechCrunch running tracker of 2026 tech layoffs where employers cited AI)
  Source: TechCrunch https://techcrunch.com/2026/07/25/the-running-list-major-tech-layoffs-in-2026-where-employers-cited-ai/
- COUNTER-EVIDENCE: Unemployment among workers most exposed to AI rose 0.77 percentage points since 2022, versus 0.85 percentage points for the least-exposed workers - i.e. AI-exposed workers fared marginally better. Employment in AI-exposed occupations remains 'fairly stable', software developer job postings have grown faster than other occupations over the past year, and firms adopting enterprise AI saw employment grow 10% in the two following years. When researchers added controls, entry-level employment declines were 'not notable until 2024'. (aggregate labour-market impact of AI Stanford Institute for Economic Policy Research (SIEPR) policy brief)
  Source: Stanford SIEPR https://siepr.stanford.edu/publications/policy-brief/what-really-happening-jobs-separating-ai-hype-reality
- Fiverr lost 21.9% of its annual active buyers year-over-year in Q2 2026, falling to 2.7 million, while revenue fell 10% to $97.8 million and FY2026 guidance was cut to $356-372 million (a 17%-14% annual decline). Spend per remaining buyer rose 15.6% to $368. (freelance marketplace buyer attrition Fiverr Q2 2026 earnings release)
  Source: StockTitan / Fiverr International https://www.stocktitan.net/news/FVRR/fiverr-announces-second-quarter-2026-o0quisjjfetp.html

## Occupations

### Data entry keyers — CRITICAL

- Category: Data & Admin
- Verdict: Disappearing. The measured decline is already steep.
- Evidence: BLS projects employment falling from 141,600 (2024) to 104,900 (2034), a -25.9% change — a loss of 36,700 jobs.
- Automated: Transcribing values from forms, invoices and scanned documents into databases — replaced by OCR plus large language model extraction that reads unstructured documents directly.
- What survives: Exception handling, validating low-confidence extractions, and regulated data domains (clinical trials, court records) where an auditable human sign-off is required.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Entry-level / new-graduate software engineers — CRITICAL

- Category: Software
- Verdict: Disappearing. The measured decline is already steep.
- Evidence: SignalFire's 2026 State of Tech Talent report finds new-grad hiring down roughly 65% at the 'Tech Majors' and about 76% at early-stage startups versus a 2019 baseline. Separately, front-end engineer share of roles fell about 25% since ChatGPT's launch while AI/ML engineer share rose 39%.
- Automated: The junior tasks that used to justify a first hire — boilerplate code, unit tests, bug triage, simple CRUD features — now produced by coding assistants under a senior engineer's supervision.
- What survives: SignalFire's read is that junior roles were cut for cost while experienced engineers became more valuable as AI leverage; the surviving entry path runs through AI/ML and research engineering rather than generalist front-end work.
- Source: SignalFire, State of Tech Talent Report 2026 (published June 22, 2026) — https://www.signalfire.com/blog/signalfire-state-of-talent-report-2026

### Switchboard operators, including answering service — CRITICAL

- Category: Customer Support
- Verdict: Disappearing. The measured decline is already steep.
- Evidence: BLS projects employment falling from 36,600 (2024) to 27,000 (2034), a -26.3% change.
- Automated: Inbound call answering and transfer, after-hours message taking — absorbed by cloud PBX systems and AI voice agents.
- What survives: Small-practice medical answering services and legal intake where callers are distressed or the intake is regulated.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Telemarketers — CRITICAL

- Category: Customer Support
- Verdict: Disappearing. The measured decline is already steep.
- Evidence: BLS projects employment falling from 67,400 (2024) to 52,500 (2034), a -22.1% change — a loss of 14,900 jobs.
- Automated: Outbound cold-call scripts, lead qualification and objection handling — now performed by synthetic-voice AI agents at a fraction of per-call cost.
- What survives: High-value complex B2B sales, relationship selling, and regulated financial products where disclosure rules require a licensed human.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Telephone operators — CRITICAL

- Category: Customer Support
- Verdict: Disappearing. The measured decline is already steep.
- Evidence: BLS projects employment falling from 4,000 (2024) to 2,900 (2034), a -27.5% change.
- Automated: Call routing, directory lookup and message relay — replaced by automated attendants, voice IVR and conversational speech recognition.
- What survives: Emergency dispatch, relay services for deaf and hard-of-hearing callers, and hospital switchboards where misrouting carries clinical risk.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Word processors and typists — CRITICAL

- Category: Data & Admin
- Verdict: Disappearing. The measured decline is already steep.
- Evidence: BLS projects employment falling from 40,000 (2024) to 25,600 (2034), a -36.1% change — the single steepest decline of any US occupation in the 2024–2034 projections.
- Automated: Manual document keying, formatting and transcription of dictated or handwritten copy — now handled by speech-to-text, autocomplete and generative drafting inside standard office software.
- What survives: Residual demand in legal, medical and government settings requiring certified accuracy and chain-of-custody on documents; survivors migrate into records management and document-control roles.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Bookkeeping, accounting and auditing clerks — ENDANGERED

- Category: Finance
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 1,613,400 (2024) to 1,519,100 (2034), a -6% change and a net loss of 94,300 positions. BLS notes 'software innovations have automated many of the tasks performed by' these workers.
- Automated: Transaction coding, ledger reconciliation and invoice matching — automated by cloud accounting platforms with AI categorisation and bank-feed matching.
- What survives: BLS expects the role to become 'more analytical and advisory'; about 170,000 annual openings persist from replacement demand.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/office-and-administrative-support/bookkeeping-accounting-and-auditing-clerks.htm

### Cashiers — ENDANGERED

- Category: Retail
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects a -10% change 2024–2034, a loss of 313,600 positions from about 3.16 million jobs. McKinsey separately projects a 630,000 decline in cashier demand by 2030. WEF ranks cashiers among the fastest-declining roles globally.
- Automated: Scanning, tendering and bagging — displaced by self-checkout, computer-vision 'just walk out' systems and the shift to online sales.
- What survives: Age-restricted sales, theft intervention, and customer-experience-led retail formats; roughly 542,600 openings a year still occur from turnover alone.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/sales/cashiers.htm

### Computer programmers — ENDANGERED

- Category: Software
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 121,200 (2024) to 113,900 (2034), a -6% change. BLS states companies are expected to 'leverage technologies, including artificial intelligence (AI), to automate repetitive programming tasks.'
- Automated: Writing code to someone else's specification — repetitive implementation work now generated by coding assistants, with higher-skilled work shifting to software developers.
- What survives: Legacy-system maintenance (COBOL, embedded firmware), and migration into the software developer role, which BLS projects growing 16%.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/computer-and-information-technology/computer-programmers.htm

### Customer service representatives — ENDANGERED

- Category: Customer Support
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 2,814,000 (2024) to 2,660,300 (2034), a -5% change and a loss of 153,700 jobs. McKinsey projects customer service and sales declining by roughly 2.0 million jobs by 2030. Stanford's Canaries research names customer service as one of two flagship case studies of entry-level displacement.
- Automated: Tier-1 ticket handling — order status, returns, password resets, FAQ answers — now resolved end-to-end by AI chat and voice agents.
- What survives: BLS notes companies keep in-house centres 'for complex inquiries such as refunding accounts or confirming insurance coverage.' Klarna publicly reversed a 700-agent AI substitution and rehired humans. Roughly 341,700 annual openings persist.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/Office-and-Administrative-Support/Customer-service-representatives.htm

### Executive secretaries and executive administrative assistants — ENDANGERED

- Category: Data & Admin
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects a -2% change 2024–2034, roughly 7,900 fewer positions; median pay $74,260 (2024). The wider secretaries and administrative assistants group is projected flat at 0%. WEF's Future of Jobs Report 2025 ranks administrative assistants among the fastest-declining roles, and McKinsey projects 710,000 fewer administrative assistant positions by 2030.
- Automated: Scheduling, correspondence drafting, travel booking, meeting notes and expense filing — increasingly handled by AI assistants and self-service tools used by the executive directly.
- What survives: Chief-of-staff style work — gatekeeping, political judgement, discretion with sensitive information, and physical office management. About 358,300 annual openings persist across the group.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/office-and-administrative-support/secretaries-and-administrative-assistants.htm

### File clerks — ENDANGERED

- Category: Data & Admin
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 84,300 (2024) to 70,900 (2034), a -15.9% change.
- Automated: Physical and digital filing, retrieval and indexing — replaced by full-text and semantic search over document stores.
- What survives: Records retention governance, legal holds, and physical archive custody in courts and healthcare.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Insurance underwriters — ENDANGERED

- Category: Finance
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects a -3% change 2024–2034, about 3,300 fewer positions from 127,000 jobs. BLS states: 'Automated underwriting software allows workers to process applications quickly, reducing the need for underwriters.'
- Automated: Risk scoring and application review against rulebooks — automated underwriting software now clears standard applications without human touch.
- What survives: Large commercial, specialty and catastrophe risk where exposure is novel and models have no precedent; model governance and pricing oversight.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/business-and-financial/insurance-underwriters.htm

### Legal secretaries and administrative assistants — ENDANGERED

- Category: Legal
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects a -6% change 2024–2034 for legal secretaries, eliminating roughly 9,000 jobs. Median pay $54,140 (2024). BLS notes AI systems and digital tools 'enable staff in many organizations to prepare their own documents without the help of secretaries.'
- Automated: Drafting standard pleadings and correspondence, calendar and docket management, document formatting — increasingly handled by AI drafting inside practice-management software.
- What survives: Court filing rules and e-filing compliance, client contact, and deadline liability management where a missed docket date is malpractice.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/office-and-administrative-support/secretaries-and-administrative-assistants.htm

### Medical transcriptionists — ENDANGERED

- Category: Healthcare Admin
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects a -5% change 2024–2034, about 2,200 fewer positions from 43,900 jobs. BLS states: 'Technological advances in speech recognition and natural language processing allow physicians to document patient encounters in real time, reducing the need for medical transcriptionists.'
- Automated: Typing up dictated clinical notes — replaced by ambient speech recognition and NLP that documents the encounter in real time.
- What survives: Editing and quality-assurance of AI-generated notes, specialty vocabularies, and accented or multi-speaker dictation; the same tools raise output per remaining worker.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/healthcare/medical-transcriptionists.htm

### News analysts, reporters and journalists — ENDANGERED

- Category: Media
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects a -4% change 2024–2034, about 1,900 fewer jobs from 49,300 positions, citing 'declining advertising revenue in radio, newspapers, and television' and consolidation among news organisations.
- Automated: Commodity news production — earnings summaries, sports recaps, aggregation and rewrites — plus the AI-driven collapse in referral traffic that funded newsrooms.
- What survives: Original reporting, source cultivation, investigative work and on-the-ground presence — none of which AI can source. Median pay $60,280 with a top decile above $162,000, indicating a widening barbell.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/media-and-communication/reporters-correspondents-and-broadcast-news-analysts.htm

### Office machine operators — ENDANGERED

- Category: Data & Admin
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 25,500 (2024) to 21,600 (2034), a -15.2% change.
- Automated: Operating copying, scanning, mailing and duplicating equipment as a dedicated function — absorbed into self-service devices and digital workflows.
- What survives: High-volume regulated mail operations (ballots, statements, legal notices) with chain-of-custody requirements.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Order clerks — ENDANGERED

- Category: Data & Admin
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 89,500 (2024) to 74,100 (2034), a -17.2% change.
- Automated: Taking, keying and confirming orders by phone, fax and email — replaced by self-service portals, EDI and AI email-to-order parsing.
- What survives: Complex configured orders, key-account management, and resolving supply exceptions.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Payroll and timekeeping clerks — ENDANGERED

- Category: Finance
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 161,100 (2024) to 134,200 (2034), a -16.7% change — a loss of 26,900 jobs.
- Automated: Timesheet collection, pay calculation, deduction reconciliation — absorbed by integrated payroll platforms with automated compliance rules.
- What survives: Multi-jurisdiction tax compliance, garnishment and benefits edge cases, and audit response.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Prepress technicians and workers — ENDANGERED

- Category: Creative
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 26,200 (2024) to 22,300 (2034), a -14.6% change.
- Automated: Manual typesetting, colour separation, imposition and proofing — automated by desktop publishing and now generative layout tools.
- What survives: Colour management for premium print, packaging compliance, and press-side troubleshooting.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Print binding and finishing workers — ENDANGERED

- Category: Media
- Verdict: Shrinking. The decline is measurable and underway.
- Evidence: BLS projects employment falling from 35,800 (2024) to 30,000 (2034), a -16.1% change.
- Automated: Not AI directly — automated finishing lines plus the collapse of print volume as content moved digital. Included as a control case for technology-driven decline.
- What survives: Short-run specialty printing, packaging and premium physical goods where craft finishing is the product.
- Source: US Bureau of Labor Statistics, Employment Projections 2024–2034 — https://www.bls.gov/emp/tables/fastest-declining-occupations.htm

### Camera operators, television, video and film — ADAPTING

- Category: Media
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 1% growth 2024–2034 for camera operators (36,400 jobs, about 400 added), noting that consolidation of roles and robotic cameras limit growth. Film and video editors fare better at 4% growth (43,500 jobs). Median pay $68,810 for camera operators.
- Automated: Static and repetitive camera work — replaced by robotic camera systems, automated tracking and drone rigs.
- What survives: Cinematography as authorship — framing, lens and lighting choices — plus documentary and live event work where the operator must anticipate an unscripted moment.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/media-and-communication/film-and-video-editors-and-camera-operators.htm

### Craft and fine artists (including illustrators) — ADAPTING

- Category: Creative
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 0% change 2024–2034 (52,000 to 52,100 jobs), with about 4,400 annual openings. Median pay $56,260, but craft artists earn only $38,480 while fine artists including illustrators earn $60,560. The WashU/NYU Upwork study found image-related freelancers lost 3.7% of monthly jobs and 9.4% of monthly earnings after image-generation tools launched.
- Automated: Commercial illustration and concept art at the commodity end — book covers, spot illustration, background and asset art — generated from prompts.
- What survives: Authorship and provenance, physical original work, gallery and collector markets, and commissions where the buyer is purchasing a specific human's vision.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/arts-and-design/craft-and-fine-artists.htm

### Editors — ADAPTING

- Category: Media
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 1% growth 2024–2034 (115,800 to 116,500 jobs), noting 'traditional print newspapers and magazines lose ground to other media formats.' Median pay $75,260; about 9,800 annual openings, mostly replacement.
- Automated: Line editing, copyediting, proofreading and style enforcement — largely automated by grammar and style AI.
- What survives: Commissioning and judgement about what is worth publishing, fact verification, legal risk review, and increasingly the verification of AI-generated copy — a growing category of editorial work.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/media-and-communication/editors.htm

### Freelance copywriters and content writers — ADAPTING

- Category: Creative
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: A study by Xiang Hui and Oren Reshef (WashU Olin) and Luofeng Zhou (NYU) using Upwork data found that after ChatGPT's launch in November 2022, writing-related freelancers saw monthly jobs decline 2% and monthly earnings decline 5.2%. Counterintuitively, more experienced and higher-skilled freelancers lost more than lower-skilled ones — Reshef: 'For higher-quality workers, instead of being protected, you're losing your competitive edge.'
- Automated: Volume content — SEO articles, product descriptions, listicles, email and ad copy — the exact briefs that made up the freelance content market.
- What survives: Original reporting and interviews, brand voice ownership, subject-matter authority, and accountability for claims. Writers who move up-market into strategy and editing of AI output fare better than those competing on word count.
- Source: Washington University Olin Business School — https://olin.washu.edu/about/news-and-media/news/2023/08/study-ai-tools-cause-a-decline-in-freelance-work-and-incomeat-least-in-the-short-run.php

### Graphic designers — ADAPTING

- Category: Creative
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 2% growth 2024–2034 (265,900 to 271,500 jobs), noting 'automated design tools, such as artificial intelligence (AI), may reduce the need for companies to contract with freelance graphic designers.' The WEF Future of Jobs Report 2025 added graphic designers to its fastest-declining list for the first time, specifically attributing it to generative AI. A WashU/NYU study of Upwork found image-related freelancers saw monthly jobs fall 3.7% and monthly earnings fall 9.4% after image-generation tools launched.
- Automated: Production design at the low end — social assets, ad variants, simple logos, stock layout work — now generated from prompts, hitting freelance and commodity briefs hardest.
- What survives: Brand strategy, art direction, design systems, and client-facing judgement about what should be made at all. About 20,000 annual openings persist. Median pay $61,300, above the $49,500 all-occupations median.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/arts-and-design/graphic-designers.htm

### Interpreters and translators — ADAPTING

- Category: Translation
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: A CEPR study using American Community Survey and Lightcast data across 696 US local labour markets found each 1 percentage point increase in machine translation usage cut translator employment growth by about 0.7 percentage points — an estimated loss of about 28,000 new translator positions that would otherwise have been created over 2010–2023. Job postings requiring Spanish fell 1.4pp in high-MT-adoption areas, Chinese about 1.3pp, German about 0.8pp. BLS projects only 2% growth 2024–2034 (75,300 to 76,600 jobs).
- Automated: Bulk written translation of documents, subtitles and product content — neural machine translation now produces publishable-quality first drafts, converting the job from translation to post-editing.
- What survives: BLS notes 'many of these jobs cannot be entirely automated because computers cannot yet produce work comparable to what human translators do in most cases.' Live interpreting, legal and medical certification, literary and marketing transcreation, and low-resource languages remain human. Median pay $59,440.
- Source: CEPR / VoxEU — https://cepr.org/voxeu/columns/lost-translation-ais-impact-translators-and-foreign-language-skills

### Loan officers — ADAPTING

- Category: Finance
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 2% growth 2024–2034 (301,400 to 306,500 jobs), attributing the slow growth to 'increased use of productivity-enhancing technology in loan processing.' Median pay $74,180.
- Automated: Application intake, document collection and preliminary credit assessment — handled by automated underwriting software that 'produces a loan recommendation based on the applicant's financial status.'
- What survives: BLS notes loan officers still review software output before final decisions. Relationship origination, commercial and non-standard lending, and regulatory accountability for adverse-action decisions remain human.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/business-and-financial/loan-officers.htm

### Paralegals and legal assistants — ADAPTING

- Category: Legal
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 0% change 2024–2034 (376,200 jobs in 2024 to 376,800 in 2034), explicitly attributing the flatline to 'advances in technology, including artificial intelligence (AI)' expected to 'make paralegals and legal assistants more efficient at tasks such as conducting research and preparing documents, which may reduce demand for these workers.'
- Automated: First-pass document review, legal research memos, discovery culling and routine document assembly — tasks where LLMs now handle volume that used to require billable paralegal hours.
- What survives: Client interviewing, court and agency filing procedure, chain-of-custody on evidence, and verification of AI output — a live professional-responsibility issue after sanctions for AI-fabricated citations. About 39,300 annual openings persist.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/legal/paralegals-and-legal-assistants.htm

### Photographers — ADAPTING

- Category: Creative
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 2% growth 2024–2034 (151,200 to 154,000 jobs), citing that 'the ease and quality of photos taken by smartphones may reduce the need for professional photographers' and that online stock services are 'possibly dampening demand.' Median pay $42,520, below the $49,500 all-occupations median.
- Automated: Stock and generic commercial imagery — now generated on demand, on top of long-running smartphone and stock-library pressure.
- What survives: Events that happen once — weddings, news, sports — plus portraiture, product shoots requiring real objects, and any use where provenance of a real photograph matters.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/media-and-communication/photographers.htm

### Receptionists — ADAPTING

- Category: Data & Admin
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 0% change 2024–2034, just 300 net new jobs. BLS states: 'Organizations continue to automate or consolidate administrative functions... many organizations use computer software, websites, mobile applications, or other technology to interact with the public or customers.' Median pay $37,230.
- Automated: Call answering, visitor check-in and appointment booking — displaced by self-service kiosks, booking apps and AI phone agents.
- What survives: Physical security and visitor screening, in-person hospitality, and de-escalation in medical and social-service settings. About 128,500 annual openings persist from turnover.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/office-and-administrative-support/receptionists.htm

### Travel agents — ADAPTING

- Category: Customer Support
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 2% growth 2024–2034 (65,700 to 67,200 jobs), stating 'the ability of travelers to use online resources to research vacations and book their own trips may limit demand for travel agents.' Median pay $48,450; about 7,100 annual openings.
- Automated: Itinerary search, fare comparison and booking — automated first by online travel agencies and now by AI trip planners that assemble complete itineraries.
- What survives: BLS notes people 'are expected to continue relying on travel agents for their advice on popular or unique destinations and for their ability to handle travel issues' — the value shifted from booking to crisis handling and curated expertise.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/sales/travel-agents.htm

### Writers and authors — ADAPTING

- Category: Creative
- Verdict: Changing, not vanishing. Tasks are being automated; the role is holding.
- Evidence: BLS projects 4% growth 2024–2034 from 135,400 jobs (about 4,900 added), citing a shift toward online media and self-publishing. Median pay $72,270, with the lowest decile under $41,080 and the top decile above $133,680.
- Automated: Draft generation and formulaic long-form content; the staff-writing tier has compressed faster than the headline number suggests.
- What survives: The pay spread signals a barbell: commodity writing collapses toward the bottom decile while named authorship, expertise and voice hold value at the top. AI cannot supply lived experience or take reputational responsibility for a claim.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/media-and-communication/writers-and-authors.htm

### Accountants and auditors — RESILIENT

- Category: Finance
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 5% growth 2024–2034 (1,579,800 to 1,652,600 jobs, 72,800 added), faster than the 3% all-occupations average, with about 124,200 annual openings. BLS: 'Some routine accounting tasks may be automated... The automation of routine tasks, such as data entry, will instead make accountants' advisory and analytical duties more prominent.' Median pay $81,680. Note the contrast: WEF's Future of Jobs Report 2025 lists accountants and auditors among globally declining roles, so the US and global pictures diverge.
- Automated: Routine bookkeeping, reconciliation and data entry within the accounting function — but this freed capacity rather than cutting headcount.
- What survives: Attestation and audit opinion carry legal liability that cannot be delegated to a model; advisory, tax strategy and controls design are growing.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/business-and-financial/accountants-and-auditors.htm

### Data scientists — RESILIENT

- Category: Data & Admin
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 34% growth 2024–2034 from 245,900 jobs (about 82,500 added), against a 3% all-occupations average, with about 23,400 annual openings. Median pay $112,590.
- Automated: Routine model fitting and exploratory analysis — automated, while demand for the role grew faster than almost any other occupation.
- What survives: Problem formulation, data quality judgement, causal reasoning and deciding whether a model should be deployed at all.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/math/data-scientists.htm

### Electricians — RESILIENT

- Category: Skilled Trades
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 9% growth 2024–2034, much faster than the 3% all-occupations average — about 77,400 new jobs from 818,700, with roughly 81,000 annual openings. Median pay $62,350, above the all-occupations median. Growth drivers include renewable energy and grid integration work.
- Automated: Nothing material — the work requires physical dexterity in unstructured, non-repeatable environments.
- What survives: Every part of it. Diagnosing faults in existing buildings, code compliance, and licensed liability. Notably, AI data centre buildout is itself a demand driver for this trade.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/construction-and-extraction/electricians.htm

### HVAC mechanics and installers — RESILIENT

- Category: Skilled Trades
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 8% growth 2024–2034, 'much faster than the average for all occupations,' with approximately 40,100 annual openings. Median pay $59,810. Entry typically requires a postsecondary nondegree award of six months to two years plus EPA refrigerant certification.
- Automated: Nothing material; diagnostic software assists but installation and repair remain manual.
- What survives: Physical installation, on-site diagnosis of systems that were never documented, and regulated refrigerant handling. Also one of the shortest training pipelines among resilient occupations.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/installation-maintenance-and-repair/heating-air-conditioning-and-refrigeration-mechanics-and-installers.htm

### Heavy and tractor-trailer truck drivers — RESILIENT

- Category: Transport
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 4% growth 2024–2034, from 2,235,100 to 2,324,400 jobs (89,300 added), with approximately 237,600 annual openings. Median pay $57,440.
- Automated: Nothing at scale yet, despite a decade of autonomous-trucking predictions. BLS's 2024–2034 projection makes no mention of autonomous vehicles affecting employment.
- What survives: First and last mile, loading and securing freight, customer interaction, and roadside problem-solving. A second useful counterexample to confident automation forecasts — though Challenger data shows transportation was the second-largest job-cut sector in 2026 (41,748 cuts YTD) for reasons other than autonomy.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/transportation-and-material-moving/heavy-and-tractor-trailer-truck-drivers.htm

### Home health and personal care aides — RESILIENT

- Category: Healthcare
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 17% growth 2024–2034, from 4,347,700 to 5,087,500 jobs — 739,800 added positions and approximately 765,800 annual openings, the largest of any occupation. Growth is driven by 'the rising number of older people' and the shift of long-term care from institutional to home settings. Median pay $34,900, well below the $49,500 all-occupations median.
- Automated: Essentially nothing — demographic demand overwhelms any automation effect.
- What survives: Physical caregiving, mobility assistance and companionship in unstructured home environments. The important caveat: this is the largest growth category in the economy and it is also among the lowest paid — resilience does not mean prosperity.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/healthcare/home-health-and-personal-care-aides.htm

### Human resources specialists — RESILIENT

- Category: Data & Admin
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 6% growth 2024–2034, faster than the 3% average, from 944,300 to 1,002,700 jobs (58,400 added). Median pay $72,910. Indeed Hiring Lab data shows AI mentions in HR job postings doubled from 4.4% to 8.8% of postings during 2025 — the role is absorbing AI rather than being absorbed by it.
- Automated: Resume screening, scheduling and candidate sourcing — automated, but the occupation still grows.
- What survives: Employee relations, investigations, compensation judgement and legal compliance — plus new work governing AI use in hiring, which is now regulated in several jurisdictions.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/business-and-financial/human-resources-specialists.htm

### Market research analysts — RESILIENT

- Category: Data & Admin
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 7% growth 2024–2034 (941,700 to 1,004,700 jobs, 63,000 added), with about 87,200 annual openings. Growth is driven by 'increasing use of data and market research across many industries.' Median pay $76,950.
- Automated: Survey tabulation, report generation and descriptive statistics — automated, but demand for analysis grew faster than the automation.
- What survives: Framing the research question, choosing what to measure, and translating findings into commercial decisions — judgement tasks that expand as the cost of analysis falls.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/business-and-financial/market-research-analysts.htm

### Plumbers, pipefitters and steamfitters — RESILIENT

- Category: Skilled Trades
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 4% growth 2024–2034, about 22,700 new jobs from 504,500, with approximately 44,000 annual openings. Median pay $62,970, above both the all-occupations median ($49,500) and the construction trades median ($56,490).
- Automated: Nothing material.
- What survives: All of it — repair work in existing structures is the definition of a non-repeatable physical task. Entry via 4–5 year apprenticeship with state licensing.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/construction-and-extraction/plumbers-pipefitters-and-steamfitters.htm

### Radiologists — RESILIENT

- Category: Healthcare
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: In 2016 Geoffrey Hinton predicted radiologists would be obsolete within five to ten years. As of 2026, average radiologist salary reached $571,000, up 9% year-on-year; the number of active US radiologists rose about 10% over the last decade; there were 7,469 open radiologist postings in May 2026 with 1,470 open 60+ days; imaging case volume rose 25% from 2018 to early 2025. Hinton has since narrowed his claim to image analysis tasks rather than the profession.
- Automated: Image pattern detection — the single most confidently predicted AI replacement in medicine, which did not happen at the occupation level.
- What survives: The canonical counterexample: AI absorbed a task (detection) inside a job that also involves clinical correlation, multidisciplinary consultation, procedures and legal responsibility for a diagnosis. Faster reads increased imaging demand rather than reducing headcount.
- Source: Fortune — https://fortune.com/article/ai-godfather-radiologists-obsolete-salaries-up-to-571k-demand-growing/

### Registered nurses — RESILIENT

- Category: Healthcare
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 5% growth 2024–2034, above the 3% all-occupations average, from about 3.4 million jobs, with approximately 189,100 job openings projected annually. Median pay $93,600.
- Automated: Documentation and charting assistance only; the core work is physical, judgement-based and relational.
- What survives: Hands-on patient assessment, procedures, escalation judgement and patient communication. AI reduces charting burden, which increases time available for care rather than reducing nurse demand.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/healthcare/registered-nurses.htm

### Software developers, QA analysts and testers — RESILIENT

- Category: Software
- Verdict: Holding up. Hard to automate, or growing.
- Evidence: BLS projects 15% growth 2024–2034 for the combined occupation (software developers alone 16%, QA analysts 10%), from 1,895,500 jobs in 2024 with 287,900 positions added and about 129,200 annual openings. BLS attributes demand to 'the continued expansion of software development for artificial intelligence (AI), Internet of Things (IoT), robotics, and other automation applications.' Median pay $133,080 for developers. Indeed Hiring Lab data shows software development postings rose about 15% since February 2025 while total postings fell 7% — but remain roughly 27.5% below pre-pandemic levels, with 71% of the recovery concentrated in senior roles.
- Automated: Boilerplate implementation and test authoring — but at the occupation level AI created more demand than it removed, even as the entry rung was pulled up.
- What survives: Architecture, system design, debugging production incidents and requirements negotiation. The occupation is resilient in aggregate but the seniority mix has shifted decisively — the risk is concentrated at the entry level, not across the profession.
- Source: US Bureau of Labor Statistics, Occupational Outlook Handbook — https://www.bls.gov/ooh/computer-and-information-technology/software-developers.htm

## Layoffs where the employer cited AI

### Microsoft — July 2026 — 4,800 roles

- Sector: Tech
- Share of workforce: 2.1%
- Attribution: explicit
- What the employer said: Chief People Officer Amy Coleman: "AI is changing how work gets done," while also stating: "I also want to be direct that the roles eliminated today are not being replaced by AI."
- Roles hit: Xbox/gaming hardest hit — 1,600 immediate cuts plus a further ~1,600 through FY2027
- Note: Microsoft simultaneously cited AI as environmental cause AND denied AI replacement — the canonical example of hedged AI attribution. Reported alongside ~$190bn annual AI infrastructure spending.
- Source: ABC News — https://abcnews.com/Business/microsoft-laying-off-4800-workers-ai-changing-work/story?id=134518435

### monday.com — July 2026 — 620 roles

- Sector: Tech / SaaS
- Share of workforce: ~20% (600-630 across reports)
- Attribution: explicit
- What the employer said: Co-CEO Eran Zinman said the restructuring is about "doing the work with AI and not just managing it," adding "We have never seen such a significant opportunity in software, driven by such exciting technology." He said the cut was "not to replace people with AI, nor to improve margins" and that the "vast majority" of savings would be reinvested into talent, products and AI.
- Roles hit: Across the company globally
- Note: Disclosed to the SEC via Form 6-K; company simultaneously lifted its 2026 margin outlook. Announcement dated July 22, 2026; TechCrunch dates its tracker entry July 25.
- Source: CIO — https://www.cio.com/article/4200330/monday-com-cuts-20-of-its-workforce-to-restructure-for-the-ai-era.html

### GitLab — June 2026 — 350 roles

- Sector: Tech / DevOps
- Share of workforce: 14%; also exiting 22 countries
- Attribution: explicit
- What the employer said: CEO Bill Staples: "Agents work at machine scale, and they're pushing competitors to the brink. This quarter we began a generational rebuild of git to support the scale and features required for 100x growth." GitLab said agentic AI workloads are straining developer infrastructure beyond its original design.
- Roles hit: Across the company; savings redirected to infrastructure rebuild for agentic workloads
- Note: $30-35m restructuring cost. AI framed as a demand-side forcing function requiring reinvestment, not as replacing the cut workers.
- Source: TechCrunch — https://techcrunch.com/2026/06/03/gitlab-cuts-14-of-staff-as-it-scales-its-platform-to-serve-ai-workloads/

### Oracle — June 2026 — 21,000 roles

- Sector: Tech
- Share of workforce: ~13% (162,000 at May 31 2025 → 141,000 at May 31 2026)
- Attribution: explicit
- What the employer said: Oracle's FY2026 annual SEC filing states: "the adoption and deployment of AI technologies across our operations have resulted, and may continue to result, in reductions to our workforce."
- Roles hit: Across the company over fiscal 2026
- Note: STRONGEST DOCUMENTARY EVIDENCE IN THE DATASET — a company stating AI-driven workforce reduction in an SEC filing. Reflects a full fiscal year of cumulative reductions disclosed at once (filed ~June 22-23, 2026), not a single announcement. Also see https://www.cnbc.com/2026/06/23/oracle-ai-job-cuts-layoffs-21000.html Figure covers Oracle's fiscal 2026 (year ended May 31, 2026), not calendar 2026: headcount fell from 162,000 to 141,000 per the annual report.
- Source: The Next Web / CNBC — https://thenextweb.com/news/oracle-21000-layoffs-ai-data-centres

### Coinbase — May 2026 — 700 roles

- Sector: Finance / Crypto
- Share of workforce: 14%
- Attribution: explicit
- What the employer said: CEO Brian Armstrong: "This is a new way of working, and we need to leverage AI across every facet of our jobs." He said engineers now "ship in days what used to take weeks," while also citing a down market: "we're currently in a down market and need to adjust our cost structure now so that we emerge from this period leaner, faster and more efficient."
- Roles hit: Broad; company moving toward "one-person teams" combining engineering, design and PM, and eliminating "pure managers" in favour of player-coaches; some roles shifted to managing fleets of AI agents
- Note: $50-60m restructuring cost; completion expected by Q2 2026. Mixed causation: AI efficiency plus crypto market downturn.
- Source: CBS News — https://www.cbsnews.com/news/coinbase-layoffs-ai-brian-armstrong/

### PayPal — May 2026 — 4,760 roles

- Sector: Finance / Fintech
- Share of workforce: ~20%
- Attribution: explicit
- What the employer said: New CEO Enrique Lores described "stripping out redundant structures and speeding up the integration of AI across the business," saying "PayPal needs to focus. We need to recommit to the fundamentals." TechCrunch's tracker records PayPal's commitment to "aggressively adopt AI" in development and customer service.
- Roles hit: Development, operations, customer service; redundant structures
- Note: Figure reported as ~4,500+ (TechCrunch) to 4,760 (Yahoo Finance); phased over 2-3 years. $1.5bn cost-savings target. Lores replaced Alex Chriss in March 2026.
- Source: Yahoo Finance — https://finance.yahoo.com/markets/stocks/articles/paypal-layoffs-ceo-cuts-20-154944985.html

### Cloudflare — May 2026 — 1,100 roles

- Sector: Tech / Internet infrastructure
- Share of workforce: 20%
- Attribution: explicit
- What the employer said: CEO Matthew Prince: "The vast majority of those we laid off last week were measurers" and "AI systems can now measure an organization with objective detail and precision that was previously impossible." He said the company retained "builders" (engineers) and "sellers."
- Roles hit: Middle management, finance, legal, internal audit, revenue recognition — roles Prince calls "measurers"
- Note: Announced after record revenue. Prince framed it as restructuring, not cost-cutting; Cloudflare's engineering headcount subsequently rose 45%. One of the clearest 'AI made this job category obsolete' statements on record.
- Source: Fortune — https://fortune.com/2026/05/21/cloudflare-ceo-matthew-prince-layoffs-ai-automation-measurers/

### General Motors — May 2026 — 600 roles

- Sector: Automotive
- Share of workforce: >10% of the IT organization (reported range 500-600; some outlets say up to 1,000 software roles)
- Attribution: implied
- What the employer said: GM: "GM is transforming its Information Technology organization to better position the company for the future." TechCrunch reported GM was replacing legacy IT staff with people who can build AI systems from the ground up rather than merely use AI tools; laid-off employees told CNBC AI played a role in the decision.
- Roles hit: Salaried IT / software; GM simultaneously hiring for AI-native development, data engineering, cloud engineering, agent and model development, and prompt engineering
- Note: GM's official statement stops short of naming AI as the cause; the AI framing comes from reporting and employee accounts. Headcount reported inconsistently across outlets (500-1,000).
- Source: TechCrunch — https://techcrunch.com/2026/05/11/gm-just-laid-off-hundreds-of-it-workers-to-hire-those-with-stronger-ai-skills/

### Cisco — May 2026 — 4,000 roles

- Sector: Tech
- Share of workforce: ~5%
- Attribution: explicit
- What the employer said: CFO Mark Patterson: "This was really not a savings-driven restructure. [It's] really realigning resources around silicon, optics, security and AI." CEO Chuck Robbins: "The companies that will win in the AI era will be those with focus, urgency, and the discipline to continuously shift investment toward the areas where demand and long-term value creation are strongest."
- Roles hit: Across the company; resources moved to silicon, optics, security and AI
- Note: Cisco's second major AI-framed restructuring (after Aug 2024).
- Source: Benzinga — https://www.benzinga.com/markets/tech/26/05/52550445/cisco-lays-off-4000-employees-ai-overhaul

### Meta — May 2026 — 8,000 roles

- Sector: Tech / Social media
- Share of workforce: ~10%; plus ~6,000 open requisitions cancelled
- Attribution: implied
- What the employer said: CONTRADICTORY. Mark Zuckerberg: "Getting everyone internally to use AI tools and getting to do the work more efficiently is not the thing that's driving layoffs." TechCrunch's tracker nonetheless lists Meta among AI-citing employers, referencing Meta's framing that "success isn't a given" in AI competition and its redeployment of ~7,000 staff into AI roles.
- Roles hit: Recruiting, sales, middle management, non-AI-adjacent product work; remaining teams reorganised into AI "pods" under Superintelligence Labs
- Note: ATTRIBUTION DISPUTED BY THE COMPANY. Announced despite record Q1 2026 revenue of $56.31bn and net income of $26.8bn, alongside $125-145bn of planned annual AI infrastructure spending. Zuckerberg promised no further company-wide cuts in 2026.
- Source: The Next Web — https://thenextweb.com/news/meta-layoffs-8000-zuckerberg-ai-reality-may-2026

### Intuit — May 2026 — 3,000 roles

- Sector: Tech / Finance software
- Share of workforce: 17% (of 18,200 as of July 2025)
- Attribution: implied
- What the employer said: CONTRADICTORY. Goodarzi's memo said the reductions aimed to "reduce complexity, simplify the company's corporate structure, and deliver better AI products." But CNBC reported the same week that Goodarzi said the 17% cut had "nothing to do with AI."
- Roles hit: Corporate structure simplification across the company
- Note: CONFLICTING PUBLIC STATEMENTS — see https://www.cnbc.com/2026/05/20/intuit-ceo-says-companys-17percent-workforce-cut-had-nothing-to-do-with-ai.html. Intuit signed AI deals with Anthropic and OpenAI around the same period. This is Intuit's second AI-framed restructuring after July 2024.
- Source: TechCrunch — https://techcrunch.com/2026/05/20/intuit-to-lay-off-over-3000-employees-to-refocus-on-ai/

### Commerzbank — May 2026 — 3,000 roles

- Sector: Finance / Banking
- Share of workforce: ~8%
- Attribution: explicit
- What the employer said: Commerzbank: "With Momentum 2030, the bank will leverage the potential of AI even more," deploying AI agents to "significantly reduce labour-intensive tasks." CEO Bettina Orlopp: "This proves that our strategy is working – and that it has more potential than originally planned."
- Roles hit: Labour-intensive process work: account switching, KYC, document checks, contract drafting; ~10% of workforce capacity freed for redeployment
- Note: Over four years, with €600m invested in AI and €500m of expected annual value from 2030. Separate from Commerzbank's Feb 2025 announcement of 3,900 cuts by 2028.
- Source: Computer Weekly — https://www.computerweekly.com/news/366642998/Commerzbank-to-cut-3000-jobs-as-it-leverages-AI-even-more

### Google (Alphabet) — May 2026 — — roles

- Sector: Tech
- Share of workforce: Undisclosed; reported range 1,500-3,000+
- Attribution: implied
- What the employer said: Per TechCrunch's tracker, cuts came via rolling performance reviews and structural reorganisation tied to AI investment priorities.
- Roles hit: Google Cloud, cybersecurity (including Threat Intelligence Group / Mandiant teams), HR
- Note: LOW CONFIDENCE ON NUMBER — Google does not disclose totals for rolling cuts, and TechCrunch itself lists the count as undisclosed. Note: Google explicitly denied AI was the cause of its earlier Jan 2024 ad-sales cuts.
- Source: TechCrunch — https://techcrunch.com/2026/07/25/the-running-list-major-tech-layoffs-in-2026-where-employers-cited-ai/

### Snap — April 2026 — 1,000 roles

- Sector: Tech / Social media
- Share of workforce: 16% (from ~5,261 at Dec 2025); 300+ open roles also closed
- Attribution: explicit
- What the employer said: CEO Evan Spiegel's memo: "rapid advancements in artificial intelligence enable our teams to reduce repetitive work, increase velocity, and better support our community, partners, and advertisers." Snap said teams are "leveraging AI tools to drive meaningful progress" across Snapchat+, ad platform performance and infrastructure.
- Roles hit: Global, across functions
- Note: Expected to cut annualized costs by >$500m by H2 2026. Snap's stock rose on the news.
- Source: TechCrunch — https://techcrunch.com/2026/04/15/snap-is-cutting-1000-jobs-16-of-its-workforce/

### IBM — April 2026 — — roles

- Sector: Tech
- Share of workforce: Reported range 3,000-9,000, undisclosed by IBM
- Attribution: implied
- What the employer said: Per TechCrunch's tracker, IBM tied the reduction to a plan to roughly triple entry-level hiring in AI and hybrid cloud.
- Roles hit: Red Hat engineering, HR
- Note: LOW CONFIDENCE ON NUMBER. IBM does not disclose layoff counts; its standard language is a "low single-digit percentage of our global workforce." Range is journalistic estimate only.
- Source: TechCrunch — https://techcrunch.com/2026/07/25/the-running-list-major-tech-layoffs-in-2026-where-employers-cited-ai/

### Atlassian — March 2026 — 1,600 roles

- Sector: Tech
- Share of workforce: ~10%
- Attribution: explicit
- What the employer said: Co-CEO Mike Cannon-Brookes: "Our approach is not 'AI replaces people.' But it would be disingenuous to pretend AI doesn't change the skill mix we need or roles required in certain areas." He said the company is "reshaping our skill mix and changing how we work to build for the future."
- Roles hit: >900 roles in software R&D; engineering and data science hit hardest. Geography: ~40% North America, ~30% Australia, ~16% India
- Note: NOTE: TechCrunch's tracker renders this quote inverted ("AI doesn't change the mix of skills we need"). The verified quote is the opposite. Announced alongside a CTO change.
- Source: The Next Web — https://thenextweb.com/news/atlassian-is-cutting-1600-jobs-and-replacing-its-cto

### Salesforce — February 2026 — — roles

- Sector: Tech
- Share of workforce: <1,000 roles, under 1%
- Attribution: explicit
- What the employer said: Per TechCrunch's tracker, Salesforce cited Agentforce benefits driving a decline in support cases and said it did not need to backfill support engineer roles. Benioff separately said on the May 27, 2026 earnings call that engineering headcount had been "mostly flat... because we have been using AI to create more efficiency for our engineers."
- Roles hit: Marketing, product, data analytics, support
- Note: Small round relative to Salesforce's 2025 support-org cut. A further ~86 California WARN-noticed roles across Agentforce, MuleSoft and Marketing Cloud followed in June 2026.
- Source: TechCrunch — https://techcrunch.com/2026/07/25/the-running-list-major-tech-layoffs-in-2026-where-employers-cited-ai/

### Block (formerly Square) — February 2026 — 4,000 roles

- Sector: Finance / Fintech
- Share of workforce: ~40-50% (from >10,000 to just under 6,000)
- Attribution: explicit
- What the employer said: CEO Jack Dorsey: the "intelligence tools we're creating and using, paired with smaller and flatter teams, are enabling a new way of working," and "Block will be built with intelligence at the core of everything we do. how we work, how we create, how we serve our customers."
- Roles hit: Across the company
- Note: Percentage reported variously as ~40% (Fortune) and ~50% (Fox Business/TechCrunch) depending on baseline headcount used. Severance: 20 weeks base + 1 week/year tenure. Dorsey predicted others would do the same "within the next year."
- Source: Fox Business — https://www.foxbusiness.com/markets/jack-dorsey-cuts-nearly-half-block-workforce-amid-major-ai-overhaul

### Amazon — January 2026 — 16,000 roles

- Sector: Tech / Retail
- Share of workforce: ~9% of corporate workforce
- Attribution: implied
- What the employer said: Per TechCrunch's AI-layoff tracker, Amazon tied the reduction to using "AI extensively" reducing the need for some jobs; Amazon's own memo from Beth Galetti emphasised "reducing layers, increasing ownership, and removing bureaucracy" and said "While many teams finalized their organizational changes in October, other teams did not complete that work until now."
- Roles hit: Corporate roles broadly, with core product and engineering notably affected
- Note: ATTRIBUTION CONTESTED. The January memo itself did not name AI; TechCrunch and CNN framed it as AI-driven based on Jassy's prior statements. Brings Amazon's total to 30,000 since Oct 2025 — the largest in company history.
- Source: GeekWire — https://www.geekwire.com/2026/amazon-confirms-16000-more-job-cuts-bringing-total-layoffs-to-30000-since-october/

### Dell Technologies — January 2026 — 11,000 roles

- Sector: Tech / Hardware
- Share of workforce: ~10% (108,000 → ~97,000)
- Attribution: implied
- What the employer said: Per TechCrunch's tracker, Dell tied the reduction to its pivot toward AI-optimized server revenue, which it projected would roughly double by fiscal 2027.
- Roles hit: Across the company
- Note: Figure comes from Dell's fiscal-2026 annual filing (filed ~March 18, 2026), covering the full fiscal year rather than a single announcement; $569m severance. TechCrunch dates the announcement Jan 30, 2026. Dell's filing language describes 'disciplined cost management' rather than naming AI as the cause.
- Source: Tech Startups (reporting Dell FY2026 10-K) — https://techstartups.com/2026/03/18/dell-lays-off-11000-employees-spends-569m-as-ai-shift-accelerates-job-cuts-filing-shows/

### Verizon — November 2025 — 13,000 roles

- Sector: Telecom
- Share of workforce: ~13% of ~99,600
- Attribution: implied
- What the employer said: Verizon framed the cut as operational efficiency. CEO Dan Schulman separately said AI already lets Verizon "resolve 85% of all issues without manual intervention," described building agent capabilities into the tech stack, and warned AI could produce unemployment "on the scale of the Great Depression within two to five years."
- Roles hit: Largely management and corporate; largest cut in company history
- Note: Verizon did not formally attribute the 13,000 cuts to AI in the announcement; Schulman's AI-and-jobs statements came around and after it. Cuts completed by April 2026.
- Source: Light Reading — https://www.lightreading.com/ai-machine-learning/ai-mad-verizon-to-continue-with-cuts-after-ceo-s-jobs-warning

### HP Inc. — November 2025 — 6,000 roles

- Sector: Tech / Hardware
- Share of workforce: ~10% (range 4,000-6,000)
- Attribution: explicit
- What the employer said: CEO Enrique Lores said the restructuring is "to drive customer satisfaction, product innovation, and productivity through artificial intelligence adoption and enablement," citing two years of piloting "agentic AI" for efficiency gains.
- Roles hit: Product development, internal operations, customer support, administrative functions
- Note: Announced on FQ4 2025 earnings call, Nov 25-27 2025. Completion by end of fiscal 2028.
- Source: Computing — https://www.computing.co.uk/news/2025/hp-6000-job-cuts-ai

### Goldman Sachs — October 2025 — — roles

- Sector: Finance / Banking
- Share of workforce: Not quantified — "limited reduction in roles"
- Attribution: explicit
- What the employer said: Internal memo from David Solomon, John Waldron and Denis Coleman: "While we are still in the early innings in terms of assessing where AI solutions can best be deployed, it's become increasingly clear that our operational efficiency goals need to reflect the gains that will come from these transformational technologies" and "For Goldman to fully benefit from the promise of AI, we need greater speed and agility in all facets of our operations."
- Roles hit: Across the firm; headcount growth to be constrained through year-end
- Note: No number disclosed. Goldman had 48,300 employees at end-Sept 2025, up 1,800 from end-2024, and expected headcount to still rise for the full year. 'OneGS 3.0' program.
- Source: Fortune (reporting Reuters-obtained memo) — https://fortune.com/2025/10/14/goldman-sachs-layoffs-headcount-earnings-ai-efficiencies

### Nestlé — October 2025 — 16,000 roles

- Sector: Consumer goods / Food
- Share of workforce: ~6%
- Attribution: implied
- What the employer said: CEO Philipp Navratil: "The world is changing, and Nestlé needs to change faster." The company tied the reduction to automation and digitalisation, including AI, across operations while targeting >CHF3bn ($3.7bn) of savings by end-2027.
- Roles hit: 12,000 white-collar roles plus 4,000 manufacturing and supply-chain roles
- Note: Nestlé did not specify which roles AI would replace. Largest non-tech automation-attributed cut of 2025. Over two years.
- Source: SupplyChainBrain — https://www.supplychainbrain.com/articles/42669-nestle-to-cut-16k-jobs-in-bid-to-ramp-up-automation

### Amazon — October 2025 — 14,000 roles

- Sector: Tech / Retail
- Share of workforce: ~4% of ~350,000 corporate staff
- Attribution: explicit
- What the employer said: SVP Beth Galetti: "This generation of AI is the most transformative technology we've seen since the Internet," saying the company must "organize more leanly, with fewer layers and more ownership, to move as quickly as possible." CEO Andy Jassy (June 2025 memo): "As we roll out more Generative AI and agents, it should change the way our work is done... we will need fewer people doing some of the jobs that are being done today."
- Roles hit: Human resources (PXT), devices and services, operations, communications, sustainability
- Note: Largest single corporate layoff in Amazon history at the time. Amazon spent $55.6bn on tech infrastructure in H1 2025.
- Source: TechCrunch — https://techcrunch.com/2025/10/28/amazon-to-cut-14000-corporate-jobs

### Chegg — October 2025 — 388 roles

- Sector: Tech / EdTech
- Share of workforce: 45%
- Attribution: explicit
- What the employer said: Chegg: "The new realities of AI and reduced traffic from Google to content publishers have led to a significant decline in Chegg's traffic and revenue."
- Roles hit: Across the company; Dan Rosensweig returned as CEO in the same announcement
- Note: Second Chegg round of 2025 (after 22% in May). Cumulative 2025 reduction well over 60% of staff. AI as demand destruction, not internal automation.
- Source: Fox Business — https://www.foxbusiness.com/technology/chegg-lay-off-45-workforce-ai-google-traffic-shifts-batter-business

### UPS — October 2025 — 48,000 roles

- Sector: Logistics
- Share of workforce: Not disclosed as a percentage
- Attribution: implied
- What the employer said: UPS executives called it the company's "most significant strategic shift in history," tied to network reconfiguration, automation and digitalisation. UPS did not name AI specifically.
- Roles hit: ~34,000 operational (warehouse/delivery) and ~14,000 management/corporate; ~93 facilities closed
- Note: WEAK AI ATTRIBUTION. Primary drivers were the Amazon volume unwind and physical warehouse automation, not generative AI. Include only as automation-driven. Cumulative figure through Sept 2025.
- Source: American Bazaar — https://americanbazaaronline.com/2025/10/29/ups-lays-off-approximately-48000-workers-469324/

### Salesforce — September 2025 — 4,000 roles

- Sector: Tech
- Share of workforce: Support org cut from 9,000 to ~5,000 (44% of that function)
- Attribution: explicit
- What the employer said: CEO Marc Benioff: "I've reduced it from 9,000 heads to about 5,000 because I need less heads" with AI. He said roughly half of all customer conversations are now handled by AI agents, and that Agentforce could return ~100 million customer leads/calls that human staff had never been able to answer.
- Roles hit: Customer support / customer service
- Note: Salesforce said it "successfully redeployed hundreds of employees" into professional services, sales and customer success. The single most-cited direct AI-for-headcount substitution of 2025.
- Source: The Register — https://www.theregister.com/2025/09/02/salesforce_4000_jobs_ai

### Accenture — September 2025 — 11,000 roles

- Sector: Consulting / Professional services
- Share of workforce: ~1.4% (791,000 → 779,000 over 3 months)
- Attribution: explicit
- What the employer said: CEO Julie Sweet: "We are exiting — on a compressed timeline — people where reskilling, based on our experience, is not a viable path for the skills we need." She added: "Every new wave of technology has a time where you have to train and retool. Accenture's core competency is to do that at scale."
- Roles hit: Staff whose skills could not be reskilled for AI work, across the firm
- Note: $865m severance charge. Accenture reported $2.6bn of AI consulting bookings over the prior six months. Program was expected to run through November 2025.
- Source: HR Grapevine (reporting Accenture Q4 FY25 earnings) — https://www.hrgrapevine.com/us/content/article/2025-09-30-accenture-cuts-11000-jobs-as-it-goes-all-in-on-ai-driven-future

### Lufthansa Group — September 2025 — 4,000 roles

- Sector: Airlines / Transport
- Share of workforce: Not disclosed as a percentage
- Attribution: explicit
- What the employer said: Lufthansa: "The profound changes brought about by digitalization and the increased use of artificial intelligence will lead to greater efficiency in many areas and processes."
- Roles hit: Administrative roles, mostly in Germany
- Note: Target date 2030, announced at Capital Markets Day. Administrative/back-office only; flight and ground ops excluded.
- Source: Skift — https://skift.com/2025/09/29/lufthansa-to-cut-4000-jobs-by-2030-as-it-looks-to-ai/

### Microsoft — July 2025 — 9,000 roles

- Sector: Tech
- Share of workforce: ~4%
- Attribution: implied
- What the employer said: No direct AI attribution from Microsoft; framed as organizational streamlining amid record AI infrastructure spending.
- Roles hit: Gaming/Xbox, sales, and across divisions
- Note: Second of Microsoft's large 2025 rounds; ~15,000 total across 2025 per year-end tallies. Attribution to AI is journalistic, not corporate.
- Source: CNBC — https://www.cnbc.com/2025/07/02/microsoft-laying-off-about-9000-employees-in-latest-round-of-cuts.html

### Recruit Holdings (Indeed, Glassdoor) — July 2025 — 1,300 roles

- Sector: Tech / HR services
- Share of workforce: ~6% of Recruit's HR-tech segment
- Attribution: explicit
- What the employer said: CEO Hisayuki "Deko" Idekoba: "AI is changing the world, and we must adapt by ensuring our product delivers truly great experiences." He added: "Hiring is still too slow and too hard, and we're using AI to make it simple and more personal — for both job seekers and employers."
- Roles hit: R&D, people operations (HR), sustainability; primarily US roles
- Note: Also reported by Bloomberg. Idekoba became CEO of the HR-tech segment as part of the same reorg.
- Source: Entrepreneur — https://www.entrepreneur.com/business-news/indeed-glassdoor-cutting-1300-employees-adapting-to-ai/494542

### Tata Consultancy Services (TCS) — July 2025 — 12,000 roles

- Sector: Tech / IT services
- Share of workforce: ~2%
- Attribution: explicit
- What the employer said: CEO K Krithivasan: "The ways of working are changing," pointing to new AI technologies, and "We have been deploying AI at scale and evaluating skills we will be requiring for the future." He said "there are roles where redeployment has not been effective."
- Roles hit: Predominantly middle- and senior-level employees where redeployment failed
- Note: TCS publicly resisted calling it an AI-replacement layoff, framing it as a skills mismatch created by AI deployment. India's MeitY and Karnataka state sought explanations.
- Source: Silicon Republic — https://www.siliconrepublic.com/business/tata-consultancy-services-tcs-layoff-2pc-12000-ai-artificial-intelligence

### CrowdStrike — May 2025 — 500 roles

- Sector: Tech / Cybersecurity
- Share of workforce: 5%
- Attribution: explicit
- What the employer said: CEO George Kurtz said "AI is a force multiplier throughout the business" and that AI "flattens our hiring curve," describing AI as "reshaping every industry."
- Roles hit: Non-customer-facing roles; continued hiring in customer-facing and product engineering
- Note: Announced despite 29% revenue growth. $36-53m in restructuring charges.
- Source: MSSP Alert — https://www.msspalert.com/news/crowdstrike-cutting-5oo-jobs-as-ai-streamlines-business

### Chegg — May 2025 — 248 roles

- Sector: Tech / EdTech
- Share of workforce: 22%
- Attribution: explicit
- What the employer said: CEO Nathan Schultz blamed Google's "implementation and expansion of AI Overviews" for collapsing traffic, and noted OpenAI's free ChatGPT Plus for college students and similar Anthropic moves as direct threats to the business.
- Roles hit: Chegg Study and corporate services hit hardest; concentrated in US and Canada
- Note: AI as demand destruction (students/search substituting for Chegg), not internal automation. Q1 2025 revenue fell 30% YoY.
- Source: TechRadar — https://www.techradar.com/pro/chegg-announces-move-to-reduce-workforce-by-22-percent-as-students-turn-to-ai

### IBM — May 2025 — — roles

- Sector: Tech
- Share of workforce: "a couple hundred" HR roles
- Attribution: explicit
- What the employer said: CEO Arvind Krishna told the WSJ that IBM's AskHR agents replaced several hundred HR staff and now handle 94% of routine HR tasks, but added: "Our total employment has actually gone up, because what [AI] does is it gives you more investment to put into other areas" — namely programmers and salespeople.
- Roles hit: Human resources
- Note: WIDELY MISREPORTED. Viral claims of '8,000 IBM HR jobs replaced by AI' are not supported by IBM's own statements; Krishna said a few hundred roles and rising total headcount. Treat the 8,000 figure as unverified.
- Source: Entrepreneur (reporting WSJ interview) — https://www.entrepreneur.com/business-news/ibm-ceo-ai-replaced-hundreds-of-human-resources-staff/491341

### Business Insider — May 2025 — — roles

- Sector: Media
- Share of workforce: 21%
- Attribution: explicit
- What the employer said: CEO Barbara Peng: "We are going all-in on AI — and we're off to a strong start," noting over 70% of staff already use Enterprise ChatGPT regularly with a goal of 100%. She cited "extreme traffic drops outside of our control" from AI search and reduced search referrals.
- Roles hit: "Touch every department"; company also exiting most of its commerce business
- Note: Absolute headcount not disclosed in the memo. Dual AI causation: AI search destroying referral traffic, plus internal AI adoption.
- Source: Nieman Journalism Lab — https://www.niemanlab.org/2025/05/business-insider-will-lay-off-21-of-staff-amid-ai-disruption-and-extreme-traffic-drops-outside-of-our-control/

### Klarna — May 2025 — — roles

- Sector: Finance / Fintech
- Share of workforce: 40% cumulative reduction restated
- Attribution: explicit
- What the employer said: Siemiatkowski told CNBC AI helped shrink the workforce ~40% and that the AI assistant does the work of 700 agents. In the same period he conceded the push went too far: "Cost unfortunately seems to have been a too predominant evaluation factor... Really investing in the quality of the human support is the way of the future."
- Roles hit: Customer service
- Note: Restatement of the same cumulative reduction as the 2024 entry — do not double-count. Notable as the highest-profile partial reversal of an AI-driven staffing cut.
- Source: CNBC — https://www.cnbc.com/2025/05/14/klarna-ceo-says-ai-helped-company-shrink-workforce-by-40percent.html

### Microsoft — May 2025 — 6,000 roles

- Sector: Tech
- Share of workforce: ~3%
- Attribution: implied
- What the employer said: Microsoft did not name AI as the cause, citing "organizational changes." Contemporaneous context: Satya Nadella said ~20-30% of code in Microsoft repos was AI-written, and the cuts landed alongside ~$80bn of FY25 AI capex.
- Roles hit: Disproportionately software engineers and middle management
- Note: Microsoft has consistently declined to attribute 2025 cuts to AI. Included because Challenger and CNBC's year-end tally counted Microsoft (~15,000 across 2025) among AI-citing employers; treat attribution as contested.
- Source: CNBC — https://www.cnbc.com/2025/07/02/microsoft-laying-off-about-9000-employees-in-latest-round-of-cuts.html

### Duolingo — April 2025 — — roles

- Sector: Tech / EdTech
- Share of workforce: Not disclosed (contractors only)
- Attribution: explicit
- What the employer said: CEO Luis von Ahn's "AI-first" memo: Duolingo will "gradually stop using contractors to do work that AI can handle." He added: "This isn't about replacing Duos with AI. It's about removing bottlenecks" and "We can't wait until the technology is 100% perfect. We'd rather move with urgency and take occasional small hits on quality than move slowly and miss the moment."
- Roles hit: Contractors doing work AI can handle
- Note: No headcount disclosed. Von Ahn later walked back the memo's framing (Aug 2025), saying full-time staff were not being laid off and the company was still hiring.
- Source: Tech Wire Asia — https://techwireasia.com/2025/04/duolingo-shifts-toward-ai-first-model-phases-out-contractor-roles/

### Workday — February 2025 — 1,750 roles

- Sector: Tech
- Share of workforce: 8.5%
- Attribution: explicit
- What the employer said: CEO Carl Eschenbach: "Companies everywhere are reimagining how work gets done, and the increasing demand for AI has the potential to drive a new era of growth for Workday." He said the company would prioritize "innovation investments like AI and platform development, and rigorously evaluating the ROI of others across the board."
- Roles hit: Broad; company continued hiring in AI/platform roles and select geographies
- Source: Engadget — https://www.engadget.com/ai/workday-lays-off-1750-employees-citing-demand-for-ai-134922398.html

### DBS Bank — February 2025 — 4,000 roles

- Sector: Finance / Banking
- Share of workforce: ~4,000 of 8,000-9,000 contract/temp staff (permanent staff 41,354 unaffected)
- Attribution: explicit
- What the employer said: Outgoing CEO Piyush Gupta said the cuts would come as AI takes over project work: "For the first time, I'm struggling to create jobs. I'm struggling to say how I will repurpose people to create jobs."
- Roles hit: Temporary and contract staff on project work; bank simultaneously adding ~1,000 AI roles
- Note: Over three years, via natural attrition as projects complete. DBS also upskilling 13,000 staff.
- Source: Fortune — https://fortune.com/asia/2025/02/25/dbs-singapore-to-cut-temp-jobs-ai-piyush-gupta/

### Klarna — December 2024 — 2,105 roles

- Sector: Finance / Fintech
- Share of workforce: ~40% (5,527 → 3,422)
- Attribution: explicit
- What the employer said: CEO Sebastian Siemiatkowski: "The truth is, the company has shrunk from about 5,000 to now almost 3,000 employees." Klarna said its OpenAI-powered assistant, launched Feb 2024, handled the work of 700 full-time customer service agents in its first month and would add ~$40m in annual profit. Reduction achieved mainly via a hiring freeze plus "natural attrition in a company like ours is 15-20% per year."
- Roles hit: Customer service most prominently; also recruitment and marketing functions
- Note: Not a mass layoff event — attrition plus hiring freeze across 2023-2024, disclosed publicly in late 2024 and again May 2025. Klarna later partly reversed course, with Siemiatkowski saying in May 2025 that "Cost unfortunately seems to have been a too predominant evaluation factor."
- Source: Yahoo Finance / Benzinga — https://finance.yahoo.com/news/klarna-reduces-workforce-40-attributes-151610371.html

### Cisco — August 2024 — 5,600 roles

- Sector: Tech
- Share of workforce: 7%
- Attribution: explicit
- What the employer said: CEO Chuck Robbins described "shifting hundreds of millions of dollars into AI, into AI networking for cloud, into AI infrastructure, silicon and cyber." CFO Scott Herren: "It's much more about finding efficiencies across the company so that we can pivot more resources… into the fastest growth areas within the company, which are pivoting more into AI, pivoting more into cloud and pivoting more into cybersecurity."
- Roles hit: Across the company; investment shifted to AI, cloud and cybersecurity
- Note: Second Cisco layoff of 2024. Up to $1bn in severance/restructuring charges. 7% of workforce ≈ 5,600 per Business Standard.
- Source: CFO Dive — https://www.cfodive.com/news/cisco-plans-cut-workforce-amid-ai-push-layoffs-tech/724378/

### Intuit — July 2024 — 1,800 roles

- Sector: Tech / Finance software
- Share of workforce: 10%
- Attribution: explicit
- What the employer said: CEO Sasan Goodarzi: "the era of AI is one of the most significant technology shifts of our lifetime"; "AI is igniting global innovation at an incredible pace, transforming every industry"; "Companies that aren't prepared to take advantage of this AI revolution will fall behind."
- Roles hit: Broad; company said it would rehire ~1,800 in engineering, product, sales, customer success and marketing
- Note: Intuit said ~1,000 of the 1,800 were performance-related; net headcount was to stay flat.
- Source: Inc. — https://www.inc.com/jason-aten/intuits-ceo-just-said-ai-is-reason-hes-laying-off-1800-employees-his-memo-is-worst-ive-seen-yet.html

### SAP — January 2024 — 8,000 roles

- Sector: Tech
- Share of workforce: 7.4% of 107,602
- Attribution: explicit
- What the employer said: SAP said it "will further increase its focus on key strategic growth areas, in particular Business AI. It also intends to transform its operational setup to capture organizational synergies, AI-driven efficiencies."
- Roles hit: Across the company; most handled via voluntary leave and internal reskilling rather than involuntary termination
- Note: €2bn program. SAP said total headcount would be roughly flat by year-end; this was a skills-mix swap toward genAI specialists.
- Source: CIO — https://www.cio.com/article/1297964/sap-announces-2-2b-restructuring-program-thatll-impact-8000-jobs.html

### Duolingo — December 2023 — — roles

- Sector: Tech / EdTech
- Share of workforce: ~10% of contractor workforce
- Attribution: explicit
- What the employer said: Duolingo said GPT is used to translate sentences, with "human experts validate that the output quality is high enough for teaching and is in accordance with CEFR standards"; it retained a smaller group of "content curators" to review AI output.
- Roles hit: Contract translators and content creators
- Note: Absolute headcount never disclosed. Duolingo disputed the word 'layoffs.' Cuts effective around Dec 15, 2023; reported Jan 2024.
- Source: TechCrunch — https://techcrunch.com/2024/01/09/duolingo-cut-10-of-its-contractor-workforce-as-the-company-embraces-ai

### IBM — May 2023 — 7,800 roles

- Sector: Tech
- Share of workforce: ~2.9% of ~270,000
- Attribution: explicit
- What the employer said: CEO Arvind Krishna told Bloomberg IBM would pause hiring for back-office roles he expected AI to replace, saying roughly 7,800 such jobs could be replaced by AI and automation over about five years.
- Roles hit: Back-office / non-customer-facing roles, notably HR
- Note: This was a hiring pause / attrition plan, not an announced layoff of 7,800. Figure is a projection, not executed cuts.
- Source: Bloomberg — https://www.bloomberg.com/news/articles/2023-05-01/ibm-to-pause-hiring-for-back-office-jobs-that-ai-could-kill

### BT Group — May 2023 — 55,000 roles

- Sector: Telecom
- Share of workforce: Up to ~42% of 130,000 (incl. contractors)
- Attribution: explicit
- What the employer said: CEO Philip Jansen said BT would be "a huge beneficiary of AI," called generative AI "a huge leap forward," and said the company could "go even further" on automation. Roughly 10,000 of the 55,000 roles were attributed to AI/automation replacing customer service and network work.
- Roles hit: Customer service, network build/maintenance; ~15,000 fibre engineers and ~10,000 maintenance roles as fibre build completes
- Note: Only ~10,000 of the 55,000 were AI-attributed; the rest tied to completing the fibre rollout. Target date 2030.
- Source: Forbes — https://www.forbes.com/sites/siladityaray/2023/05/18/telecoms-giant-bt-will-cut-55000-jobs-by-2030-and-replace-thousands-of-roles-with-ai/

### Dropbox — April 2023 — 500 roles

- Sector: Tech
- Share of workforce: 16%
- Attribution: explicit
- What the employer said: CEO Drew Houston: "the AI era of computing has finally arrived" and "We've believed for many years that AI will give us new superpowers and completely transform knowledge work."
- Roles hit: Across the company; company remained profitable at the time
- Note: Houston cited slowing growth AND the AI era; AI framed as reason to reallocate talent.
- Source: TechCrunch — https://techcrunch.com/2023/04/27/dropbox-lays-off-500-employees-16-of-staff-ceo-says-due-to-slowing-growth-and-the-era-of-ai/

## Companies killed or damaged by AI

### ClickUp blog (content marketing asset) — DEAD

- What it did: SEO content engine for a project-management SaaS; ~1.2M monthly organic visitors at peak
- Timeline: ? – 2025-2026
- Cause of death: Google AI Overviews / AI Mode absorbing informational queries
- Epitaph: Not a company but a business model: the SEO content blog as customer-acquisition channel. ClickUp's went from 1.19 million monthly organic visitors to under 29,000 in fifteen months - a 97.6% wipeout of an asset that cost years and millions to build.
- Peak: 1.19 million monthly organic visitors in January 2025
- Trough: fewer than 29,000 monthly visitors in April 2026
- Decline: 97.6% over 15 months
- The case against this entry: Some of this is plausibly Google penalising high-volume programmatic SEO content independently of AI Overviews. The figure is reported by PPC Land citing third-party SEO tooling, not by ClickUp.
- Source: PPC Land — https://ppc.land/ai-overviews-killed-overfishing-org-and-its-not-alone/

### Giant Freakin Robot — DEAD

- What it did: Independent science-fiction, gaming and entertainment news site
- Timeline: ? – 2024
- Cause of death: Google search algorithm shift toward large brands (pre-dating AI Overviews' dominance)
- Epitaph: From 20 million monthly uniques to a few thousand. The shutdown post was titled 'The End Of Independent Publishing.' A total-loss case, and the most extreme traffic collapse-to-closure in this dataset.
- Peak traffic: 20 million unique visitors per month
- Traffic at shutdown: a few thousand per month
- Shutdown: announced November 2024
- The case against this entry: Explicitly attributed by its founder to Google prioritising large brands in core updates, NOT to AI Overviews. It is frequently miscited as an AI casualty. Include it as evidence of the pre-AI algorithmic squeeze that AI search then compounded.
- Source: Performance Marketing World — https://www.performancemarketingworld.com/article/1894979/publisher-giant-freakin-robot-shuts-down-due-massive-drop-google-driven-traffic
- Source: Hello Partner — https://hellopartner.com/2024/11/05/news-outlet-giant-freakin-robot-shuts-down-amidst-google-troubles/

### Quartz — DEAD

- What it did: Digital business news publication founded by The Atlantic's parent company
- Timeline: 2012 – 2024-2025
- Cause of death: Owner-driven replacement of journalists with AI-generated content, followed by sale as a traffic asset
- Epitaph: Founder Zach Seward's verdict: 'Quartz is now a zombie brand, which is the most cynical move in media.' The publication was not out-competed by AI - it was hollowed out and refilled with it, keeping the domain authority and discarding the journalists.
- Sale: Sold by G/O Media to Redbrick, a Canadian software company, April 2025
- Staff: Entire editorial staff fired on sale; had kept ~80 employees when G/O acquired it in 2022
- Ownership history: David Bradley -> Uzabase -> Zach Seward (bought back for $1) -> G/O Media (2022) -> Redbrick (2025)
- Content: G/O filled the site with 2000s-era slideshows and AI-generated earnings stories
- The case against this entry: Quartz is a case of AI as instrument rather than AI as cause. Its decline began with the 2018 Uzabase sale and G/O Media's cost-stripping, both pre-ChatGPT. AI-generated earnings stories were the method of asset-stripping, not the market force that made it necessary.
- Source: Nieman Journalism Lab — https://www.niemanlab.org/2025/04/quartz-is-now-a-zombie-brand/
- Source: Axios — https://www.axios.com/2025/04/04/quartz-sold-redbrick-go-media

### Semrush — DEAD

- What it did: SEO and online-visibility analytics platform
- Timeline: ? – 2025-2026
- Cause of death: Not disruption - absorbed because AI search made 'brand visibility in LLMs' (GEO) a category Adobe wanted to own
- Epitaph: The independent SEO industry's flagship public company was absorbed into Adobe on the thesis that 'brand visibility is being reshaped by generative AI' - and that measuring how you appear inside ChatGPT and Gemini is now a bigger business than measuring how you rank on Google.
- Acquisition: Adobe agreed to acquire Semrush for approximately $1.9B at $12.00 per share, announced November 19, 2025; completed April 2026
- Growth at acquisition: 33% YoY annual recurring revenue growth in its enterprise customer segment
- The case against this entry: Explicitly NOT a death-by-AI. Semrush was growing 33% YoY in enterprise ARR when acquired. Included because it marks the end of standalone SEO tooling as a category, and because it is frequently miscast in AI-disruption narratives as a casualty rather than a beneficiary.
- Source: Adobe Newsroom — https://news.adobe.com/news/2025/11/adobe-to-acquire-semrush
- Source: Adobe Newsroom — https://news.adobe.com/news/2026/04/adobe-completes-semrush-acquisition

### TV Time (Whip Media) — DEAD

- What it did: TV and film episode-tracking app with a large consumer community
- Timeline: ? – 2026
- Cause of death: Parent company reallocating to an AI enterprise product (Helix); free consumer app deemed unsustainable
- Epitaph: Twenty-five million users switched off so the parent company could chase AI enterprise revenue. The stated reason: 'it was no longer sustainable to continue operating the service as a free app, and there was not enough demand for a paid app.'
- Shutdown date: July 15, 2026
- Users: over 25 million users; 26.4 million lifetime app installs
- Owner: Whip Media, acquired by Blue Torch Capital in early 2025
- The case against this entry: Not disrupted by AI - deprioritised because of AI. The product's user base was intact; capital simply moved to a more fashionable line of business. This pattern (consumer app sunset to fund an AI pivot) is a distinct and under-counted category of AI-related product death.
- Source: TechCrunch — https://techcrunch.com/2026/07/02/popular-tv-tracking-app-tv-time-is-shutting-down-as-company-focuses-on-ai/

### Udemy — DEAD

- What it did: Open marketplace for user-created online video courses
- Timeline: 2010 – 2025-2026
- Cause of death: Consolidation driven by AI reshaping the skills market - contested as a direct cause
- Epitaph: Udemy ceased to exist as an independent company in a $2.5 billion all-stock merger creating a single MOOC survivor. Both parties framed it as a response to AI reshaping which skills people need - not as a rescue.
- Acquisition: All-stock merger with Coursera valued at approximately $2.5B, announced December 17, 2025; expected to complete in the second half of 2026
- Market reaction: both companies reported revenue growth in Q3 2025 yet their share prices fell
- The case against this entry: Weak AI-causation case, included for honesty. TechCrunch reports both companies were growing revenue at the time of the deal. The AI framing in the announcement is strategic positioning (job postings requiring AI skills, AI-powered microlearning launches), not a distress narrative. This is edtech consolidation in a market where investors lost confidence, with AI as the stated rationale.
- Source: TechCrunch — https://techcrunch.com/2025/12/17/coursera-and-udemy-enter-a-merger-agreement-valued-at-around-2-5b
- Source: Higher Ed Dive — https://www.highereddive.com/news/coursera-to-acquire-udemy-to-create-25b-mooc-giant/808212/

### Upday (Axel Springer / Samsung) — DEAD

- What it did: News aggregation app pre-installed on Samsung phones across 34 European countries
- Timeline: 2015 – 2023
- Cause of death: Replaced by an AI-generated 'trend news generator' version of itself
- Epitaph: One of the earliest cases of a newsroom being replaced by a model rather than out-competed by one. Axel Springer shut down Upday's human-curated news operation and relaunched it as an AI trend generator, cutting staff from 150 to fewer than 70 in the process.
- Closure: closed in its existing form at the end of 2023; AI-driven relaunch planned for summer 2024
- Staff: 70 remaining employees at closure, down from 150 originally; jobs eliminated in the transformation
- Footprint: 34 European countries since 2015
- The case against this entry: Early (2023) and small. Upday's underlying problem was a weak Samsung distribution deal and poor engagement; the AI relaunch was cost-cutting justified by a technology narrative.
- Source: Reuters via Yahoo Tech — https://tech.yahoo.com/ai/articles/axel-springer-close-upday-shift-124126203.html

### overfishing.org — DEAD

- What it did: Volunteer-run educational site on fisheries depletion, ~300 pages of teaching materials and fish guides
- Timeline: 2003 – 2024
- Cause of death: Google AI Overviews summarising its content without sending visitors
- Epitaph: A 21-year-old volunteer education project reduced from 750 to 50 daily readers. Its operator shut it down rather than keep paying hosting for an audience Google now serves directly.
- Traffic before: approximately 750 unique daily visitors
- Traffic after: roughly 50 daily visitors when AI Overviews were active
- Shutdown announced: February 25, 2024
- Content: ~300 pages
- The case against this entry: Timing problem worth flagging: PPC Land dates the shutdown announcement to February 25, 2024, which precedes Google's general US launch of AI Overviews in May 2024 (though Search Generative Experience had been in Labs since May 2023). The 750-to-50 traffic figure and the AI Overviews attribution come from the site operator via PPC Land's June 2026 report. Treat the specific causal attribution to AI Overviews as plausible but not cleanly established.
- Source: PPC Land — https://ppc.land/ai-overviews-killed-overfishing-org-and-its-not-alone/

### Appen — DYING

- What it did: Human data-labelling and annotation for AI training, using a ~1 million-person contractor crowd
- Timeline: 1996 – 2023-2024
- Cause of death: Big Tech in-housing annotation and shifting to synthetic/model-generated labels; Google terminated its Appen contract in January 2024
- Epitaph: The company that labelled the data that trained the models was made redundant by the models. Losing Google - one of five clients that together made up the overwhelming majority of revenue - removed roughly a fifth of the business in a single stroke.
- H1 fy2024 revenue: US$113.4M, down 18.4% YoY
- Ex google revenue: down only 1.5% YoY, isolating the Google contract as the primary driver
- Global services division: US$63.6M, down 36.5%; excluding Google, US$50.8M, down 20.1%
- New markets division: US$49.8M, up 28.3%
- Net loss: US$17.8M in H1 FY2024, improved from a US$43.3M loss
- Crowd size: ~1 million contractors, 500+ languages, 100+ countries
- The case against this entry: A widely-shared claim that Appen fell from a A$4.3B valuation in 2020 to under $130M originates from a social-media post, not a filing, and is not verified here. Appen's problem is also client concentration and price competition from Scale AI and offshore rivals, not purely AI substitution - its 'New Markets' division grew 28.3% in the same period.
- Source: Staffing Industry Analysts — https://www.staffingindustry.com/news/global-daily-news/appen-h1-operating-revenue-falls-with-google-contract-termination

### Chegg — DYING

- What it did: US homework-help and textbook-rental subscription service for college students
- Timeline: 2005 – 2023-2026
- Cause of death: ChatGPT (Nov 2022) for direct answer substitution; Google AI Overviews (May 2024) for search referral collapse
- Epitaph: The single cleanest case of a public company gutted by a chatbot. Chegg sold instant homework answers for $19.95/month; ChatGPT gave them away, then Google AI Overviews stopped students from ever reaching the site. Revenue fell from $776M to a $43-44M quarterly run-rate guide, and $11.6B of market cap became $168M. Chegg sued Google over AI Overviews in February 2025 and cut 45% of staff eight months later.
- Peak revenue fy2021: $776.3M (+20% YoY)
- Peak subscribers: 8.2M Chegg Services subscribers at end of 2022 (7.8M at end of 2021)
- Fy2025 revenue: $376.9M, down 39% YoY; Q4 2025 $72.7M, down 49% YoY
- Fy2025 net loss: $103.4M
- Q1 2026 revenue: $63.3M, down 48% YoY
- Q2 2026 revenue: $51.8M, down 51% YoY (from $105.1M in Q2 2025)
- Q3 2026 guidance: $43-44M total net revenues
- Market cap peak: $11.64B (Dec 31, 2020)
- Market cap now: $167.94M at $1.50/share (May 28, 2026)
- Market value erased: $14.5B (Gizmodo, Nov 2024)
- Layoffs: 441 employees (~25% of workforce) Nov 2024; 388 employees (~45%) announced Oct 2025, targeting up to $110M in savings
- Subscriber loss 2024: lost half a million paid subscribers (Gizmodo)
- Student survey: 30% of students intended to use Chegg (down from 38%); 62% planned to use ChatGPT (up from 43%)
- The case against this entry: Chegg's own CEO Nathan Schultz explicitly tied the two causes together: 'we would not need to review strategic alternatives if Google hadn't launched AI Overviews.' A secondary, less-cited factor is post-pandemic normalization of college enrollment and a crackdown on academic-integrity use cases. The remaining Skilling segment ($17.5M in Q2 2026, +2%) is roughly flat, not growing into the hole.
- Source: Chegg Investor Relations / Business Wire — https://www.businesswire.com/news/home/20260806916599/en/Chegg-Reports-Second-Quarter-2026-Earnings
- Source: Chegg Investor Relations / Business Wire — https://www.businesswire.com/news/home/20260209620934/en/Chegg-Reports-2025-Fourth-Quarter-and-Full-Year-Financial-Results
- Source: Chegg Investor Relations — https://investor.chegg.com/Press-Releases/press-release-details/2026/Chegg-Reports-First-Quarter-2026-Earnings/default.aspx
- Source: Chegg Investor Relations — https://investor.chegg.com/Press-Releases/press-release-details/2023/Chegg-Reports-2022-Earnings-and-Gives-2023-Guidance/default.aspx

### Concentrix — DYING

- What it did: Customer-experience outsourcing / contact-centre operator (~440,000 employees globally)
- Timeline: ? – 2024-2026
- Cause of death: Agentic AI customer-service deployments; enterprise clients reallocating CX budget to their own AI builds
- Epitaph: The June 30, 2026 guidance cut was the day the market decided call-centre outsourcing was a melting ice cube. CEO Chris Caldwell said clients were 'reallocating budget toward their own AI investment needs rather than maintaining or expanding CX outsourcing spend.' RBC's Karl Green noted the harder fact: 'some clients have simply withdrawn customer support altogether in some areas.'
- Q2 fy2026 revenue: $2.46B, up 1.9% YoY, below the $2.47B estimate
- Guidance cut: FY2026 revenue cut to $9.93-10.03B (prior midpoint ~$10.11B); FY EPS cut to $10.83-11.18 from $11.48-12.07; Q3 EPS guide $2.65-2.77 vs $3.09 Street
- Guidance dollar cut: ~$130M off the midpoint of 2026 revenue outlook
- Stock reaction: -21.5% pre-market on June 30, 2026; -25% intraday per Bloomberg-syndicated coverage
- From 52-week high: ~-68% from a 52-week high of $62.14
- Restructuring: $175M in restructuring costs across 2026
- Market cap peak: $9.33B (Dec 31, 2021)
- Market cap now: $1.54B at $25.19/share (Aug 4, 2026)
- The case against this entry: Concentrix's revenue is still growing slightly (+1.9%). The collapse is in the multiple, not the top line - the market has repriced the business as terminally disrupted before the disruption shows up in the P&L. Offshoring to lower-cost markets and general enterprise cost-cutting are also explicitly cited alongside AI.
- Source: Investing.com — https://www.investing.com/news/stock-market-news/why-is-concentrix-stock-plunging-today-93CH-4767092
- Source: Outsource Accelerator — https://news.outsourceaccelerator.com/tp-concentrix-customer-pullback/
- Source: StockAnalysis.com — https://stockanalysis.com/stocks/cnxc/market-cap/

### Fiverr — DYING

- What it did: Marketplace for fixed-price freelance 'gigs' - logos, voiceovers, articles, video edits
- Timeline: 2010 – 2025-2026
- Cause of death: Generative AI substituting for exactly the transactional, low-value creative tasks Fiverr was built on
- Epitaph: Fiverr lost 21.9% of its buyers in a single year. The company's own framing - AI 'absorbs high-volume, low-value, transactional tasks' - is an admission that the original Fiverr business, the $5 gig, no longer has customers. What remains is a smaller, higher-value marketplace with 2.7 million buyers instead of 3.5 million.
- Q2 2026 revenue: $97.8M, down 10.0% YoY
- Active buyers: 2.7M, down 21.9% YoY
- Spend per buyer: $368, up 15.6% YoY
- Guidance: FY2026 cut to $356-372M, a 17%-14% annual decline; adjusted EBITDA $52-62M
- Net income: $4.5M GAAP ($0.12/share) in Q2 2026
- Stock reaction: -20.53% on July 29, 2026, on 5.5x average volume
- Market cap peak: $4.16B (Dec 31, 2021, year-end basis)
- Market cap now: $323.92M at $9.01/share (Aug 14, 2026)
- The case against this entry: Spend per buyer rose 15.6%, and the company remains GAAP-profitable and free-cash-flow positive. This is a deliberate managed shrink upmarket, not an uncontrolled collapse - but the buyer number is the most brutal single AI-attributable metric in this dataset.
- Source: StockTitan (Fiverr Q2 2026 release) — https://www.stocktitan.net/news/FVRR/fiverr-announces-second-quarter-2026-o0quisjjfetp.html
- Source: StockAnalysis.com — https://stockanalysis.com/stocks/fvrr/market-cap/

### Forrester Research — DYING

- What it did: Technology and market research subscriptions, consulting and events
- Timeline: 1983 – 2024-2026
- Cause of death: LLM research synthesis plus B2B budget compression toward AI spending
- Epitaph: A 40-year-old research institution now worth about a quarter of its annual revenue. Contract value, consulting and events are all shrinking simultaneously, and management is guiding to a further 9-13% revenue decline in 2026 - the profile of a business in structural, not cyclical, retreat.
- Fy2025 revenue: $396.9M, down 8%
- Q4 2025 revenue: $101.1M, down 6% YoY
- Contract value: down 6% to $292.4M
- Consulting: down 16% in Q4, down 9% for the full year
- Events: down 29% for the year
- Fy2026 guidance: $345-360M, a projected 9-13% decline
- Market cap: ~$105M, down from ~$300M a year prior; shares trading $5.50-6.50
- Valuation: ~0.25x revenue
- The case against this entry: Forrester's own AI product ('AI Access') generated $5M+ in bookings since September and reportedly shortened sales cycles by nearly 50%, so AI is also a partial remedy. The events collapse (-29%) is not plausibly AI-caused and points to broad B2B marketing budget cuts.
- Source: SaaStr — https://www.saastr.com/gartner-down-71-forrester-worth-105m-what-that-tells-us-about-b2b-spend-right-now/

### Getty Images — DYING

- What it did: Premium stock photography, editorial photojournalism and video licensing
- Timeline: 1995 – 2023-2026
- Cause of death: Generative image models compressing the microstock market (iStock), plus AI-generated search answers destroying referral and affiliate traffic
- Epitaph: A 30-year-old visual-content institution trading at 29 cents a share with $2.1 billion of debt and $51.6 million of cash. Its CEO named the mechanism directly: 'search engine referral traffic declines and the knock-on impact to our affiliate traffic sources as the search engines implement AI generated answers.' Its merger lifeboat sank in June 2026 and liquidity is now 'our immediate priority.'
- Q2 2026 revenue: $229.1M, down 2.5% YoY
- Creative revenue: $127.4M, down 2.6%
- Editorial revenue: $96.5M, up 9.2%
- Paid downloads: 19.4M
- Annual subscribers: 240,000, down from 321,000 a year earlier
- Subscription revenue retention: 88.4%, down from 93.4%
- Total debt: $2.1B as of June 30, 2026
- Cash: $51.6M, down $45M sequentially; free cash flow negative $122.6M
- Guidance: 2026 guidance withdrawn; Guggenheim Securities engaged July 2026 for 'strategic financing alternatives'
- Stock reaction: -15.2% after Q2 2026 earnings
- Market cap now: $120.58M at $0.29/share (Aug 12, 2026)
- The case against this entry: Getty's editorial business - live news, sports, the FIFA World Cup, archives - grew 9.2% and is genuinely AI-resistant, because models cannot generate a photograph of an event that happened yesterday. The damage is concentrated in commodity creative stills and iStock e-commerce. Getty's balance sheet distress is also substantially a legacy of its 2022 SPAC-era leverage, not only AI.
- Source: The Motley Fool (Q2 2026 earnings call transcript) — https://www.fool.com/earnings/call-transcripts/2026/08/17/getty-images-gety-q2-2026-earnings-call-transcript/
- Source: StockStory — https://stockstory.org/us/stocks/nyse/gety/news/earnings/getty-images-nysegety-misses-q2-cy2026-revenue-estimates-stock-drops-152percent
- Source: StockAnalysis.com — https://stockanalysis.com/stocks/gety/market-cap/

### HouseFresh — DYING

- What it did: Independent air-purifier review site doing original in-house product testing
- Timeline: ? – 2024
- Cause of death: Google's March 2024 core update prioritising large media brands and Shopping listings (not AI Overviews)
- Epitaph: A site that bought and tested air purifiers in a real lab lost 91% of its search traffic to sites that did not. Its documentation of the collapse became the canonical case study of the independent-publisher squeeze.
- Search traffic loss: 91% of search traffic lost
- Daily google visitors: fell from approximately 4,000 to 200
- Date: most severe decline on March 9, 2024
- Example: for 'best budget air purifiers' it fell from #2 since May 2023 to buried beneath 64 Google Shopping listings
- The case against this entry: Frequently and incorrectly filed under 'killed by AI.' HouseFresh's own analysis blames Google's March 2024 core update, big-media 'keyword swarming' and Shopping-listing expansion. AI-generated commerce content from large publishers is cited as a contributing factor, but AI Overviews are not the mechanism.
- Source: HouseFresh — https://housefresh.com/how-google-decimated-housefresh/

### LivePerson — DYING

- What it did: Pioneer of website live-chat and, later, rules-based enterprise 'conversational commerce' chatbots
- Timeline: 1995 – 2022-2026
- Cause of death: LLM-native customer-service agents (OpenAI-powered assistants, Agentforce-class products) making pre-LLM intent-based chatbots obsolete
- Epitaph: The company that invented the little chat bubble in the corner of your screen agreed to sell itself to an AI voice company for about $43 million in stock — less than 1% of its 2020 value — with bondholders taking a haircut so common shareholders would get anything at all. Its enterprise customers did not churn to a competitor; they churned to an LLM. The deal had not yet closed as of August 2026.
- Fy2025 revenue: $243.7M, down 22% YoY
- Fy2026 guidance: $195-207M, a further 15-20% contraction
- Q2 2026 revenue: $52.5M, down 11.9% YoY; net loss $72.56M vs $15.71M a year earlier
- Fy2025 net loss: $67.2M GAAP
- Stockholders equity: negative $51.5M as of March 31, 2026
- Net revenue retention: 78% in 2025, down from 82%, against a 105-115% target
- Exit: Agreed to be acquired by SoundHound AI, announced April 21, 2026: ~$3.33/share in SoundHound stock, ~$43M implied equity value (~$250M enterprise value including discounted debt), 22% premium to 30-day VWAP. Restated two-step merger agreement signed July 2, 2026; stockholder vote August 20, 2026; closing expected December 5, 2026. Not yet closed.
- Debt: ~$350M in secured notes; noteholders agreed to a substantial discount, without which common equity would likely have been wiped out
- Market cap peak: $4.16B (Dec 31, 2020)
- Market cap now: $23.62M at $1.94/share (July 1, 2026), a ~99.4% decline
- The case against this entry: LivePerson's collapse is over-determined: it carried heavy debt from acquisitions, had a CEO turnover and executed a botched cloud migration, all of which predate the LLM wave. But the retention number - 78% against a 105-115% target - is the cleanest evidence that customers were leaving for AI-native alternatives rather than simply shrinking.
- Source: StockTitan (SEC Form 425 filing) — https://www.stocktitan.net/sec-filings/LPSN/425-liveperson-inc-business-combination-communication-cd06c1929112.html
- Source: TradingView (10-Q summary) — https://www.tradingview.com/news/tradingview:5a0ea6ff4bac4:0
- Source: StockAnalysis.com — https://stockanalysis.com/stocks/lpsn/market-cap/

### Nerdy / Varsity Tutors — DYING

- What it did: Online live tutoring marketplace and school-district tutoring contracts
- Timeline: 2007 – 2024-2026
- Cause of death: Mixed - AI tutoring substitution is the market narrative; the documented causes are expiring federal relief funds and lack of efficacy evidence
- Epitaph: Varsity Tutors for Schools shut down in August 2026 after reaching 500 districts and $28M of peak revenue. Its 'Live+AI' hybrid product did not save it. The company is retreating to direct-to-parent tutoring with 29,100 members and a cut guidance.
- Q2 2026 revenue: $43.3M, down 4% YoY from $45.3M
- Active learning memberships: 29,100, down 5% YoY
- Guidance cut: FY2026 cut to $168-175M from $180-190M
- Net loss: $6.9M in Q2 2026, improved from $12.0M
- Business exits: First Tutors (UK) and Varsity Tutors for Schools both being wound down, $2-4M exit costs
- Vt for schools scale: served over 1,000 districts and schools; peak revenue of $28M+ in 2023
- Engagement: only 15% of students hit their recommended tutoring dosage
- The case against this entry: This is an important corrective case. Experts quoted by The 74 attribute the shutdown to the expiration of pandemic-era ESSER funds, districts demanding efficacy evidence that never materialised, and 15% dosage compliance - not to AI. Nerdy's own filings cite *higher* AI costs as a margin drag. Anyone listing this as 'killed by AI' is overreaching.
- Source: StockTitan (Nerdy Q2 2026 release) — https://www.stocktitan.net/news/NRDY/nerdy-announces-second-quarter-2026-financial-qwxebyhte9rp.html
- Source: The 74 — https://www.the74million.org/article/major-virtual-tutoring-provider-shuts-down-experts-cite-lack-of-evidence/

### RWS Holdings — DYING

- What it did: UK-listed translation, localization and IP language services group
- Timeline: ? – 2025-2026
- Cause of death: Machine translation and LLM-based localization collapsing per-word pricing
- Epitaph: The largest publicly-listed language services company lost 40% of its value in one day. Revenue barely moved - it was margin that collapsed, which is what commoditization looks like: the same words translated for less money.
- Profit warning: April 24, 2025 - shares fell over 40% in a day
- Guidance: FY adjusted pre-tax profit cut to £60-70M from £106.7M the prior year
- H1 profit: adjusted pre-tax profit ~£17M vs £46M a year earlier, down 63%
- H1 revenue: £344M vs £350.3M, down 1.8%; organic growth at constant currency +1.3%
- Margins: gross margins ~300 basis points lower YoY
- Fy25: profits declined 43% (Investing.com)
- The case against this entry: RWS frames this as a deliberate, costly transition to AI-enabled delivery rather than demand destruction, and notes three of four divisions grew. Non-trading items (FX losses, amortisation, a business disposal) accounted for £23M of the H1 decline. The AI story here is about price per unit, not volume.
- Source: MultiLingual — https://multilingual.com/rws-holdings-shares-plunge-amid-profit-warning-and-operational-challenges/
- Source: Investing.com — https://www.investing.com/news/company-news/rws-holdings-fy25-slides-reveal-ai-transformation-strategy-as-profits-decline-43-93CH-4402810

### Shutterstock — DYING

- What it did: Stock photography, video and music marketplace
- Timeline: 2003 – 2023-2026
- Cause of death: Generative image models (DALL-E, Midjourney, Stable Diffusion, Adobe Firefly) plus AI-driven search-referral collapse hitting e-commerce acquisition
- Epitaph: Shutterstock spent 2023-2025 selling its own archive to AI labs as training data, then watched those models eat the market for the archive. Down to 993,000 subscribers, 104 million paid downloads and a $200M market cap, its escape hatch - a merger with Getty - was blocked by UK regulators in June 2026, leaving it standalone and shrinking.
- Q1 2026 revenue: $199.2M, down 18% YoY from $242.6M
- Content segment: $178.1M, down 12%
- Data distribution segment: $21.0M, down 47%
- Subscribers: 993,000, down from 1,079,000 a year earlier
- Paid downloads: 104.1M, down from 120.9M
- Net loss: $47.6M vs net income of $18.7M a year earlier
- Adjusted ebitda: $42.7M, down 33% from $63.4M
- Market cap peak: $4.05B (Dec 31, 2021)
- Market cap now: $199.89M at $5.41/share (Aug 12, 2026), a ~95% decline
- Failed merger: $3.7B Getty Images merger terminated June 30, 2026 after the UK CMA required divestiture of Shutterstock's editorial business
- The case against this entry: Part of the Q1 2026 Data segment drop (-47%) is timing of metadata licence revenue recognition, not demand collapse. Revenue per customer actually rose to $284 from $244, meaning the business is losing volume customers rather than value customers. Shutterstock also books real revenue from AI data licensing, so 'AI killed it' and 'AI pays its bills' are simultaneously true.
- Source: Shutterstock Investor Relations — https://investor.shutterstock.com/news-releases/news-release-details/shutterstock-reports-first-quarter-2026-financial-results
- Source: Reuters via Investing.com — https://www.investing.com/news/stock-market-news/getty-images-scraps-shutterstock-merger-4769105
- Source: StockAnalysis.com — https://stockanalysis.com/stocks/sstk/market-cap/

### Stack Overflow — DYING

- What it did: Question-and-answer site for programmers; the default place developers went for coding help
- Timeline: 2008 – 2023-2026
- Cause of death: ChatGPT and LLM coding assistants (GitHub Copilot, Claude, Gemini) answering coding questions privately and without moderation friction
- Epitaph: The internet's programming reference library emptied out. Prosus paid $1.7B in 2021 and has written down $1.2B of it. Monthly question volume has gone from 207,204 at its 2014 peak to 1,442 in July 2026 — a 99.3% collapse — while 84% of developers now ask a model instead, a model trained largely on Stack Overflow's own archive.
- Monthly questions peak: 207,204 in March 2014
- Monthly questions now: ~3,314 in December 2025 and 1,442 in July 2026 — 99.3% below peak; 2025 totalled 108,981 questions for the whole year
- Acquisition price: $1.7B by Prosus, August 2021
- Impairments: $246M (Mar 2022) + $560M (Mar 2023) + $372M (2024) = $1.2B total written off
- Revenue: $54M (2022), $94M (2023), $98M (2024)
- Trading losses: $34M (2022), $84M (2023), $57M (2024); $175M cumulative since acquisition
- Layoffs: 28% of workforce, October 2023
- Developer AI adoption: 84% of developers use AI tools; 51% of professional developers use AI daily (Stack Overflow 2025 Developer Survey, 49,000+ respondents)
- The case against this entry: The decline in question volume began around 2014-2016, well before ChatGPT, driven by moderation strictness, duplicate-closing culture and the fact that most common questions were already answered. ChatGPT sharply accelerated an existing trend rather than starting it. The site itself is not dead: revenue was $98M in 2024 and Stack Overflow for Teams continues to sell.
- Source: PPC Land — https://ppc.land/stack-overflow-drops-to-1-442-questions-in-july-down-99-from-2014-peak/
- Source: PPC Land — https://ppc.land/stack-overflow-traffic-collapses-as-ai-tools-reshape-how-developers-code/
- Source: MyBroadband — https://mybroadband.co.za/news/business/541929-prosus-r21-billion-stack-overflow-disaster.html
- Source: VentureBeat — https://venturebeat.com/programming-development/stack-overflow-confirms-layoffs-affecting-28-of-workforce

### Teleperformance (TP) — DYING

- What it did: World's largest customer-experience outsourcer / call-centre operator
- Timeline: 1978 – 2024-2026
- Cause of death: OpenAI-powered customer service assistants; the Klarna announcement of Feb 2024 was the specific trigger event
- Epitaph: A €10 billion-revenue company was revalued by a single blog post from a Swedish buy-now-pay-later firm. Teleperformance has never recovered from February 28, 2024; by mid-2026 hedge funds were shorting the sector and Bloomberg reported investors calling call-centre stocks 'uninvestible.'
- Feb 2024 crash: Shares fell 19% to €107.9 on Feb 28, 2024, down as much as 29.3% intraday to €94.28 - a seven-year low - the day after Klarna said its OpenAI assistant handled two-thirds of service chats, equivalent to 700 full-time agents
- Market cap at crash: €4.12B on revenue of €10.21B
- Peak to trough: Down over 75% from a peak of approximately €400 in December 2021 to below €90
- July 2026: Fell as much as 16% to its lowest level in over a decade after Concentrix's guidance cut
- The case against this entry: TP's revenue has not collapsed the way its share price has - the H1 2026 declines were driven substantially by FX. This is the clearest example in the dataset of a stock priced for AI extinction while the underlying business is still large and cash-generative. Whether the market or the company is right is genuinely unresolved.
- Source: Stockopedia — https://www.stockopedia.com/share-prices/teleperformance-se-EPA:TEP/news/teleperformance-shares-plunge-on-ai-disruption-concerns-updated-019e6d7a-2d9b-73a4-820d-074c1415bac5/
- Source: Outsource Accelerator — https://news.outsourceaccelerator.com/tp-concentrix-customer-pullback/
- Source: Bloomberg (via Yahoo Finance) — https://finance.yahoo.com/technology/ai/articles/call-center-stocks-fall-worry-111631312.html

### The Planet D — DYING

- What it did: Independent travel blog run by Dave Bouskill and Debra Corbeil
- Timeline: ? – 2024-2026
- Cause of death: Google AI Overviews answering travel queries directly
- Epitaph: One of the best-known independent travel blogs, reduced to a rounding error of its former traffic. Travel is the single most AI-Overview-exposed content vertical: 'best time to visit', 'what to pack', 'is X safe' are exactly what a model answers without a click.
- Decline: 50% initial traffic loss, followed by a further 90% decline (AdExchanger); reported separately as a 90% traffic reduction following the AI Overviews rollout
- The case against this entry: The two reported figures (50%-then-90% vs a flat 90%) are not consistent with each other across sources; both come from operator self-reporting rather than measured panel data.
- Source: AdExchanger — https://www.adexchanger.com/publishers/the-ai-search-reckoning-is-dismantling-open-web-traffic-and-publishers-may-never-recover/
- Source: PPC Land — https://ppc.land/ai-overviews-killed-overfishing-org-and-its-not-alone/

### Upwork — DYING

- What it did: Global freelance marketplace for writing, design, development and admin work
- Timeline: 2015 – 2025-2026
- Cause of death: AI automation of low-end freelance tasks (copywriting, data entry, basic design) plus AI search reducing the SEO traffic that fed client acquisition
- Epitaph: CEO Hayden Brown named three causes at once in August 2026: 'a faster pace of AI-related automation, continued weakness in the labor market and further deterioration in Google SEO.' The bottom of the marketplace - the $30 blog post, the logo tweak, the data-entry gig - is being absorbed by models faster than the top of the marketplace can grow.
- Q2 2026 revenue: $191.7M, down 2% YoY
- Gsv: $966.4M, down ~4%
- Active clients: 763,000, down 4%
- Guidance cut: FY2026 cut from $760-790M to $730-750M
- Stock reaction: -21% after hours on Aug 12, 2026; had closed at $8.39, down ~15%
- AI exposed share: management estimates ~10% of GSV faces adverse AI exposure
- AI upside: AI-related work is the fastest-growing category at ~$330M annualized, up 22%; AI consulting up ~50%
- Market cap peak: $4.38B (Dec 31, 2021)
- Market cap now: $1.06B at $8.59/share (May 28, 2026)
- The case against this entry: Upwork is the strongest 'both things are true' case: AI is destroying its low-end supply while creating its fastest-growing demand category. Management's own 10%-of-GSV exposure estimate is far smaller than the stock's derating implies, and revenue is down only 2%.
- Source: The Next Web — https://thenextweb.com/news/upwork-q2-2026-ai-disruption-guidance-cut
- Source: StockAnalysis.com — https://stockanalysis.com/stocks/upwk/market-cap/

### Adobe — AT RISK

- What it did: Creative software (Photoshop, Illustrator, Premiere) and digital experience platform
- Timeline: 1982 – 2025-2026
- Cause of death: Market narrative of generative-AI substitution - DALL-E, Midjourney, Canva, Figma, Google Imagen/Gemini
- Epitaph: The most valuable creative-software franchise in history now trades at a 13x P/E while growing revenue 12%. Adobe has not been disrupted; it has been *narratively* disrupted. Whether that gap closes upward or downward is the largest open question in this dataset.
- Stock 2026: down 37% year-to-date as of June 11, 2026, trading near $222.60 against a 52-week low of $220.17
- Valuation: trailing-twelve-month P/E of approximately 13x
- Q1 fy2026 revenue: $6.40B, up 12% YoY
- Non gaap eps: $6.06, beat estimates
- Total arr: $26.06B
- Earnings record: beat earnings five consecutive quarters, each time with a negative market reaction
- The case against this entry: Explicitly a contested case. Adobe's fundamentals are growing double digits and it has beaten earnings five quarters running. There is currently no revenue or ARR evidence of AI substitution. The stock decline reflects multiple compression on a forward-looking thesis, not observed business damage. Anyone citing Adobe as an AI casualty is citing a share price, not a business.
- Source: The Motley Fool via Yahoo Finance — https://finance.yahoo.com/markets/stocks/articles/adobe-now-down-37-2026-164456666.html
- Source: The Motley Fool — https://www.fool.com/investing/2026/06/05/is-ai-going-to-bring-the-adobe-era-to-an-end/

### Business Insider — AT RISK

- What it did: Digital business and tech news publisher (Axel Springer)
- Timeline: 2007 – 2024-2025
- Cause of death: Google AI Overviews and AI search reducing organic referral traffic
- Epitaph: CEO Barbara Peng cited 'extreme traffic drops outside of our control' while cutting one in five staff - and simultaneously announced the company was 'going all-in on AI.' Business Insider lost more than half its organic search traffic in three years.
- Layoffs: 21% of staff, announced May 29, 2025, across every department
- Traffic exposure: 70% of the business has 'some degree of traffic sensitivity'
- Organic search loss: 55% of organic search traffic lost between April 2022 and April 2025
- AI adoption target: goal of 100% employee adoption of Enterprise ChatGPT, from 70% at the time
- The case against this entry: The 55% figure spans April 2022 to April 2025, meaning the majority of the decline predates AI Overviews' May 2024 US launch and overlaps with Google's Helpful Content updates and Facebook's deprioritization of news links. Business Insider also notes it now earns twice the revenue per visit it did two years earlier.
- Source: Nieman Journalism Lab — https://www.niemanlab.org/2025/05/business-insider-will-lay-off-21-of-staff-amid-ai-disruption-and-extreme-traffic-drops-outside-of-our-control/
- Source: AdExchanger — https://www.adexchanger.com/publishers/the-ai-search-reckoning-is-dismantling-open-web-traffic-and-publishers-may-never-recover/

### Duolingo — AT RISK

- What it did: Gamified language-learning app
- Timeline: 2011 – 2025-2026
- Cause of death: Market narrative of chatbot-based language practice and lower barriers to building a competing app
- Epitaph: A company growing revenue 35% year over year lost roughly four-fifths of its market value on the theory that a chatbot can teach Spanish. The first sequential MAU decline in its history was enough to break the story.
- Feb 2026 decline: stock fell 25% during February 2026, with the sharp move on the February 27 earnings report
- Peak to trough: reported as down 81% from its high as of March 2026 (Motley Fool headline)
- Q4 2025 MAU: 135.3 million falling to 133.1 million sequentially
- Q4 2025 dau: 50.5 million rising to 52.7 million
- Q4 2025 revenue: $282.9 million, up 35% YoY
- Guidance miss: Q1 revenue guide $288.5M vs $291.2M consensus; FY guide $1.2-1.22B vs $1.26B consensus
- The case against this entry: Heavily contested. Revenue grew 35% and daily active users actually increased. The bear case is a hypothesis about a future substitute, not a measured loss of customers. Duolingo also cut contract translators in favour of AI in 2024, making it simultaneously an AI adopter, an AI-disruption target, and a growth company. The '81%' figure is drawn from an article headline rather than a fetched price series - treat as approximate.
- Source: The Motley Fool — https://www.fool.com/investing/2026/03/03/why-duolingo-stock-fell-24-in-february/
- Source: The Motley Fool — https://www.fool.com/investing/2026/03/11/1-growth-stock-down-81-regret-not-buying-dip-march/

### Gartner — AT RISK

- What it did: IT research and advisory subscriptions - the industry standard 'what should we buy' service for enterprise IT
- Timeline: 1979 – 2024-2026
- Cause of death: LLMs performing on-demand research synthesis that enterprises previously paid Gartner subscriptions for
- Epitaph: Gartner's core product is a human analyst summarising a market. Contract value growth fell to 0.8% and 2026 guidance points to outright revenue decline - the first time the flywheel has run backwards. The market took $33 billion off the valuation on the thesis that a subscription to a model is cheaper than a subscription to an analyst.
- Stock decline 2025: -47.9% during calendar 2025
- Peak to trough: -71% from a November 2024 peak of $552 to ~$155
- Market cap: from over $45B to roughly $12B
- Fy2025 revenue: $6.5B, up 4%
- Q4 2025: revenue $1.8B, up 2.2%; consulting down 12.8% YoY; adjusted EPS $3.94, down 27.8%
- Contract value: Global Contract Value up 0.8% to $5.2B (up 4% excluding US federal)
- Fy2026 guidance: $6.455B - below 2025 actual revenue
- The case against this entry: Contested. US federal government cost-cutting (DOGE-era contract cancellations) is an independently large driver - excluding US federal, contract value grew 4%. Gartner's revenue is still growing. The 'AI kills Gartner' thesis is a forward-looking market bet, not yet an observed revenue phenomenon.
- Source: SaaStr — https://www.saastr.com/gartner-down-71-forrester-worth-105m-what-that-tells-us-about-b2b-spend-right-now/
- Source: The Motley Fool — https://www.fool.com/investing/2026/01/15/why-gartner-stock-was-cut-in-half-in-2025

### HuffPost — AT RISK

- What it did: High-volume digital news publisher built on search and social scale
- Timeline: 2005 – 2024-2025
- Cause of death: AI Overviews and search referral decline
- Epitaph: Half its search referrals gone. HuffPost was the archetype of the scale-and-search publisher; that model no longer has a distribution channel.
- Search referral loss: 50%
- The case against this entry: Single reported figure without a stated time window; treat as directional. HuffPost's traffic troubles also date to Facebook's 2018 news deprioritization, long before generative AI.
- Source: AdExchanger — https://www.adexchanger.com/publishers/the-ai-search-reckoning-is-dismantling-open-web-traffic-and-publishers-may-never-recover/

### Jasper (Jasper AI) — AT RISK

- What it did: AI copywriting tool built on OpenAI's API; early generative-AI breakout product
- Timeline: ? – 2023
- Cause of death: ChatGPT - the free consumer product from its own underlying model provider
- Epitaph: Jasper was killed by its own supplier. It sold a $49/month wrapper around GPT-3; then OpenAI shipped the same capability free at ChatGPT.com. The canonical 'thin wrapper' cautionary tale - and the reason 'is this a feature or a company?' became a standard VC question.
- Peak valuation: $1.5B (Series A, 2022)
- Cut valuation: approximately $1.2B, a 20% reduction
- Arr forecast: 2023 annual recurring revenue forecast cut by at least 30%
- Layoffs: July 2023
- Leadership: founder Dave Rogenmoser replaced as CEO by Timothy Young, September 2023
- The case against this entry: Jasper did not die - it pivoted to enterprise marketing teams and remains operating in 2026. The 20% internal valuation cut is modest compared with the public-market carnage elsewhere in this dataset. It is a survival-by-pivot story dressed up in the press as a death.
- Source: Maginative — https://www.maginative.com/article/jasper-cuts-internal-valuation-as-ai-growth-slows/
- Source: The Information — https://www.theinformation.com/articles/jasper-an-early-generative-ai-winner-cuts-internal-valuation-as-growth-slows

### Pearson — AT RISK

- What it did: Global education publisher, assessment and certification group
- Timeline: ? – ?
- Cause of death: None demonstrated - included as a counter-example
- Epitaph: The company most often predicted to be destroyed by AI-generated learning content grew profit 14% in H1 2026. Included as a deliberate control case.
- H1 2026: profit up 14%, with an AI upskilling strategic focus
- The case against this entry: Explicit counter-example. Pearson has repositioned toward assessment, certification and enterprise upskilling - businesses where the value is credentialing and verification, which models cannot supply. The lesson across this dataset is that AI destroys content-supply businesses and spares verification, licensing and rights businesses.
- Source: Investing.com — https://www.investing.com/news/company-news/pearson-h1-2026-slides-profit-up-14-ai-upskilling-focus-93CH-4827062
- Source: The Armchair Trader — https://www.thearmchairtrader.com/uk-shares/pearson-results-ai-represents-both-opportunity-and-risk-in-2026/

### People Inc. (formerly Dotdash Meredith) — AT RISK

- What it did: Largest US digital and print publisher - People, Investopedia, Allrecipes, Verywell, Byrdie
- Timeline: ? – 2024-2026
- Cause of death: Google AI Overviews and AI Mode intercepting the informational queries its brands were built to answer
- Epitaph: Google search traffic down 40% year over year, and CEO Neil Vogel's summary of the bind: 'If we were to turn off AI, we would turn off search.' People Inc. has responded by rebuilding the company around everything that isn't a Google session - email, social, licensing, events - which now generates 43% of digital revenue.
- Google search traffic: down 40% YoY in Q2 2026
- Core sessions: down 22% YoY
- Google share of traffic: fell from 25% of total traffic to 21% in a single quarter
- Non session revenue: up 16% YoY from $108M to $125M; rose from 39% to 43% of total digital revenue
- Aio keyword coverage: AI Overviews presence on People Inc. keywords rose from 35% to 55% of keywords between Q1 and Q2
- The case against this entry: This is the best-documented publisher case because People Inc. discloses the numbers voluntarily and is suing Google over it. Notably, the company grew non-session revenue 16% and saw ad rates rise, so total revenue has not fallen in proportion to traffic. Traffic collapse and revenue collapse are not the same event.
- Source: Relevant Audience (reporting Digiday-sourced Q2 2026 figures) — https://www.relevantaudience.com/geo/people-inc-google-search-traffic-down-40-percent/
- Source: AdExchanger — https://www.adexchanger.com/publishers/the-ai-search-reckoning-is-dismantling-open-web-traffic-and-publishers-may-never-recover/

### Stereogum — AT RISK

- What it did: Independent music criticism and news site
- Timeline: ? – 2024-2025
- Cause of death: AI search plus platform link deprioritization
- Epitaph: A 70% collapse in advertising revenue. Founder Scott Lapatine was careful about attribution: 'Facebook and X's deprioritization of links hurt, too.'
- Ad revenue decline: 70%
- The case against this entry: Lapatine himself names social-platform link suppression alongside AI search. This is a multi-cause decline, and the operator says so.
- Source: AdExchanger — https://www.adexchanger.com/publishers/the-ai-search-reckoning-is-dismantling-open-web-traffic-and-publishers-may-never-recover/

### TaskUs — AT RISK

- What it did: Outsourced customer experience, content moderation and AI data-annotation services
- Timeline: ? – 2024-2025
- Cause of death: Contested - taken private amid the broader AI repricing of BPO, but the deal documents do not attribute it to AI
- Epitaph: Removed from public markets at 1.29x sales by its own founders and Blackstone, at the exact moment public investors decided customer-experience outsourcing was uninvestable.
- Take private: $16.50 per share all-cash, $1.62B total equity value, announced May 9, 2025
- Premium: 14.74% over the last closing price
- Valuation multiple: 1.29x sales
- Buyers: co-founders together with Blackstone
- The case against this entry: Important caveat: the InsideArbitrage analysis of the deal explicitly does not attribute the transaction to AI factors, and the take-private carried a 14.74% premium rather than a distressed discount. Trade coverage framed it as happening 'amid AI shift', which is context rather than causation. TaskUs also derives revenue from AI data services - it is partly a beneficiary.
- Source: InsideArbitrage — https://www.insidearbitrage.com/2025/05/taskus-to-be-taken-private-by-co-founders-and-blackstone-for-1-62-billion/
- Source: Nearshore Americas — https://nearshoreamericas.com/taskus-to-go-private-in-1-62b-deal-amid-ai-shift/

### The New York Times — AT RISK

- What it did: National newspaper and digital subscription publisher
- Timeline: ? – ?
- Cause of death: None demonstrated - included as a counter-example
- Epitaph: The NYT lost search traffic share over the same period as everyone else - 44% to 37% - and added 1.4 million subscribers anyway. Direct audience relationships proved to be the immune system.
- Subscribers: added 1.4 million digital-only subscribers in 2025, reaching 12.78 million total, with double-digit YoY growth in both digital subscriptions and ad revenue
- Search share: search traffic share fell from 44% to 37% between 2022 and 2025
- The case against this entry: Deliberate counter-example. The publisher-extinction narrative is real for search-dependent publishers and largely false for destination brands with direct subscription relationships. Any dataset of AI casualties that omits this is misleading.
- Source: eMarketer — https://www.emarketer.com/content/washington-post-cuts-30--ai-erodes-search-traffic
- Source: AdExchanger — https://www.adexchanger.com/publishers/the-ai-search-reckoning-is-dismantling-open-web-traffic-and-publishers-may-never-recover/

### The Washington Post — AT RISK

- What it did: National American newspaper and digital news publisher
- Timeline: 1877 – 2024-2026
- Cause of death: Generative AI search discovery collapsing search-driven traffic (contributing cause among several)
- Epitaph: Marty Baron called it 'among the darkest days in the history of one of the world's greatest news organizations.' A third of the newsroom gone, with a near-halving of search-driven traffic over three years as one stated cause.
- Layoffs: Approximately one-third of the newsroom, more than 300 journalists, announced February 4, 2026
- Cumulative newsroom reduction: 400 people reduced from the newsroom over three years
- Traffic: a nearly 50% drop in search-driven traffic over the past three years, partly attributed to generative AI's impact on discovery
- Cuts: standalone sports and books sections closed, Post Reports daily podcast ended, foreign/business/national/local teams cut, 2026 Winter Olympics coverage cancelled
- Contrast: The New York Times added 1.4 million digital-only subscribers in 2025, reaching 12.78 million total
- The case against this entry: Strongly contested attribution. The Post's crisis is at least as much about owner Jeff Bezos's opinion-page interventions and the resulting subscriber cancellations, plus long-running losses, as about AI. CNBC/Reuters and NBC News coverage of the layoffs did not cite AI at all; the AI-traffic framing comes from eMarketer's analysis and leadership's stated rationale. The New York Times growing strongly over the same period is direct evidence that AI search is not uniformly fatal to publishers.
- Source: eMarketer — https://www.emarketer.com/content/washington-post-cuts-30--ai-erodes-search-traffic
- Source: NBC News — https://www.nbcnews.com/business/media/washington-post-layoffs-sports-rcna257354
- Source: CNBC / Reuters — https://www.cnbc.com/2026/02/04/washington-post-starts-massive-layoff-gutting-sports-and-foreign-coverage-reuters.html

### Wikipedia — AT RISK

- What it did: Free crowd-written encyclopedia; the reference layer of the open web
- Timeline: 2001 – 2025-2026
- Cause of death: Search engines answering directly with generative AI (built substantially on Wikipedia's own content) and social video displacing web browsing
- Epitaph: The Wikimedia Foundation stated it plainly: 'search engines are increasingly using generative AI to provide answers directly to searchers rather than linking to sites like ours.' The encyclopedia that trained the models is losing the readers - and therefore the donors and editors - who sustain it.
- Human pageviews: approximately 8% decline in human pageviews, May-August 2025 versus the same months in 2024
- The case against this entry: 8% is a modest decline by the standards of this dataset, and Wikimedia attributes it jointly to AI and to younger users shifting to social video. Wikipedia is also not commercially dependent on pageviews the way publishers are; the risk is to its contributor and donor pipeline, which is a slower and less measurable effect.
- Source: Wikimedia Foundation (Diff) — https://diff.wikimedia.org/2025/10/17/new-user-trends-on-wikipedia/

### Yext — AT RISK

- What it did: Digital knowledge / business listings management - keeping business data accurate across search engines and directories
- Timeline: ? – 2023-2026
- Cause of death: AI search changing how consumers find local business information, eroding the listings-syndication value proposition
- Epitaph: Three consecutive years of organic revenue decline, masked by acquisitions. Yext's 10-K acknowledges the shift: 'The digital consumer journey continues to change with the expansion of artificial intelligence and large language models.'
- Fy2026 revenue: up 6% headline, but 'without the Hearsay acquisition revenue would have declined 2%'
- Prior organic declines: -4% (FY2025), -1% (FY2024)
- Net income: $37.9M in FY2026
- Accumulated deficit: $669.2M
- Capital return: modified Dutch auction self-tender reduced to $140M from an initial $180M; long-term debt rising from ~$98M to ~$147.5M
- The case against this entry: Mild case. Yext is profitable ($37.9M net income) and organic decline is only -2%. A buyout process was explored and abandoned. It belongs in 'structurally threatened', not 'collapsing'.
- Source: StockTitan (Yext FY2026 10-K) — https://www.stocktitan.net/sec-filings/YEXT/10-k-yext-inc-files-annual-report-99f7d75f23bb.html