8,000 Meta workers wake up jobless; Bezos calls AI a bulldozer, not a threat; Pew says workers unimpressed with AI; Upwork fires humans; pharma layoffs soar; fewest S&P jobs since 2016

UK firms cut 8% labor after AI adoption; 35% of Americans are "disposable"; Connecticut mandates AI disclosure; China bans AI firings, Canada thinking about it; H-1B visa workers face 60-day clock

A translator in Brighton, England has watched her corporate press release work disappear. A London cinematographer is retraining as an outdoor instructor for minimum wage. These are not Americans, but the pattern behind their situations is the same one reshaping the U.S. labor market. A Morgan Stanley report published in late May [https://www.france24.com/en/live-news/20260527-my-job-is-going-uk-workers-squeezed-out-by-ai] found that British companies that adopted AI cut their workforces by 8% in the year ending October 2025. That rate exceeded comparable figures for Germany, Japan, and Australia. Listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/the-ai-labor-report/id1896663061] The IMF estimates that more than two-thirds of British workers perform tasks AI can potentially carry out. Services make up 80% of the UK economy. That combination makes the UK the most AI-exposed major economy outside the United States. The workers disappearing from its labor market are the same kinds of workers disappearing in America. AI Is a Cover Story. An MIT Professor Put a Number on It. A Fortune piece published May 31 [https://fortune.com/2026/05/31/tech-companies-ai-washing-layoffs-wix-block-snap-atlassian-disposable-workers/] quotes Paul Osterman, a professor of human resources management at MIT Sloan and the author of a book literally titled Disposable Workers. Osterman’s argument is that companies have used technology as a cover story for workforce reduction for 20 years. AI is simply the most convenient version of that cover story available right now. Disposable Workers are people whose employment relationship gives them no meaningful protection against restructuring decisions. When Cisco announced 4,000 cuts last month, its stock jumped 13%. The market rewarded the announcement, which tells you what the market thinks the layoffs are actually about. Osterman estimates that 35% of the American workforce now qualifies as “disposable workers.” That 35% figure is not a projection. It describes the current state of employment for roughly 58 million Americans. BUY NOW! [https://wimdodson.gumroad.com/l/gods_in_the_machine?_gl=1*19da5a0*_ga*MTEwMjE2MDkwMy4xNzc2OTY4MDQy*_ga_6LJN6D94N6*czE3NzY5NjgwNDEkbzEkZzEkdDE3NzY5NjgxNzIkajU5JGwwJGgw] Get the NEW Book that exposes the Narratives Tech uses to build its AI Empire. $4.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required.3.5-hr reading time. Two States Just Put Employers on Notice Connecticut Governor Ned Lamont signed SB 5 on May 30 [https://news.bloomberglaw.com/daily-labor-report/connecticuts-lamont-signs-ai-law-with-employer-notice-mandate], making Connecticut one of the first states in the country to require employers to disclose in their WARN Act filings whether a mass layoff is related to AI or other automation. The requirement takes effect October 1, 2026. Starting in October 2027, employers must give workers written notice before using AI tools that play a “substantial role” in hiring, promotion, or termination decisions. Violations are enforceable by the state attorney general. Harvard’s OnLabor noted on June 1 [https://onlabor.org/june-1-2026/] that the law puts Connecticut alongside California, Colorado, and Illinois as the only states with meaningful AI employment protections on the books. No federal equivalent exists. Instead, the Trump administration has attempted to block similar state laws through threatened loss of federal funding and litigation. The four-state patchwork is the full extent of American worker protection against AI-driven employment decisions. The EU’s equivalent rules take effect August 2. American workers employed by multinationals operating in Europe may soon have more legal protection at their European offices than at their U.S. desks. China Made It Illegal. Canada Like the Idea. Courts in Hangzhou and Beijing ruled in late April [https://www.caixinglobal.com/2026-04-30/chinese-courts-rule-companies-cannot-fire-workers-simply-to-replace-them-with-ai-102439602.html] that companies cannot terminate employees simply because AI can perform the same job at lower cost. The anchor case involved a quality assurance supervisor whose employer tried to cut his monthly pay from 25,000 to 15,000 yuan after AI automated his role. When he refused, the company fired him. The court ruled the termination illegal. The reasoning matters: AI adoption is a strategic business choice, not an unforeseeable event. Chinese labor law allows contract changes for unforeseeable disruptions. A planned technology decision does not qualify. Canadian legal analysts and the Global News outlet noted [https://globalnews.ca/news/11840683/ai-china-layoffs-court-ruling-canada/] that the Chinese ruling reignited debate about in Canada about whether Western governments are failing workers in this area. The honest answer for Canada and the U.S. is yes. The Chinese motivation may be social stability as much as worker welfare. Nevertheless, the ruling’s effect on workers is real either way. The H-1B Clock Is Running The displacement wave at Oracle, Cognizant, and Amazon hits Indian tech workers in the United States with a second consequence that American outlets are mostly missing. The AI/Labor Report is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Oracle’s restructuring cut an estimated 12,000 roles [https://www.storyboard18.com/photos/brand-marketing/ai-layoffs-2026-amazon-meta-oracle-cisco-among-tech-firms-cutting-jobs-98910.htm] from teams based in India or staffed by Indian nationals in the U.S. Cognizant cut between 12,000 and 15,000 positions to shift toward automated delivery models. For workers on H-1B visas, a layoff triggers a 60-day window to find a new employer willing to sponsor them. Miss the window, and the visa expires. These workers face AI displacement and immigration jeopardy simultaneously. They fall outside most U.S. labor displacement data. Their situation is the clearest current example of a workforce segment absorbing the full cost of the AI transition with no policy buffer in place. BUY NOW [https://wimdodson.gumroad.com/l/invasion_ai?_gl=1*g1iey2*_ga*OTc1NjU2NTcyLjE3NzQwMzA3NjM.*_ga_6LJN6D94N6*czE3NzQwMzMyMjUkbzIkZzEkdDE3NzQwMzQ2ODgkajYwJGwwJGgw]! Get the book that examines how the AI invasion already happened. You just weren’t invited. $9.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required. 2-hr reading time. Get full access to The AI/Labor Report at ailabor.substack.com/subscribe [https://ailabor.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

Jobs vanish without layoff announcements; Intuit fires the very workers who built it and buy it; 142,000 tech jobs gone; 6 million clerical workers can't adapt; they're already using the tools

The Jobs That Quietly Disappear A Yale analysis [https://insights.som.yale.edu/insights/the-real-job-destruction-from-ai-is-hitting-before-careers-can-start]finds that the biggest labor market impact of agentic AI will not appear in layoff data. It will instead appear in opportunities that quietly disappear before anyone notices. Entry-level roles that companies stop posting. Hiring pipelines that close without explanation. First jobs that never get created. The researchers describe the average firm’s response to AI not as mass layoffs but as silently closing the door to new workers. The distinction matters because the second version produces no accountability. There is no date, no memo, no severance package, and no WARN filing. There is only a graduating class that finds the door closed when it arrives. For instance, C.H. Robinson announced it is handling 29% more freight volume today than it did in early 2019 [https://fortune.com/2026/04/29/ai-agentic-entry-level-jobs-disappearing-yale-celi-sonnenfeld/] while employing 30% fewer workers. AI agents book roughly half of its carrier bookings. The company did not announce a mass layoff. It simply stopped needing the same number of people to do the same amount of work. The jobs did not disappear dramatically. They evaporated. The Company That Cut 17% of Its Workforce While Selling AI to Its Own Customers Intuit cut 3,000 workers, representing 17% of its global headcount, on May 20. [https://www.cbsnews.com/news/ai-layoffs-hiring-entry-level-workers/] The company makes TurboTax and QuickBooks. Its products serve tens of millions of American small businesses and independent filers. The workers who built and maintained those tools are now gone. The stated reason for the layoff is a shift toward AI. The practical result is that the same company selling AI-powered accounting tools to small business owners just eliminated the human workforce that built those tools. Cutting the people while selling the product to the customers those people served describes how the displacement cycle closes the loop. The small business owner using QuickBooks is the customer. The QuickBooks employee who lost her job is also the customer. The AI product is now serving both of them while employing fewer of the people who made it possible. BUY NOW! [https://wimdodson.gumroad.com/l/gods_in_the_machine?_gl=1*19da5a0*_ga*MTEwMjE2MDkwMy4xNzc2OTY4MDQy*_ga_6LJN6D94N6*czE3NzY5NjgwNDEkbzEkZzEkdDE3NzY5NjgxNzIkajU5JGwwJGgw] Get the NEW Book that exposes the Narratives Tech uses to build its AI Empire. $4.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required.3.5-hr reading time. The Layoffs That Are Not Making Business Sense CIO Magazine [https://www.cio.com/article/4171054/ai-driven-layoffs-arent-making-business-sense.html] published an analysis this month examining whether AI-driven layoffs are producing the results companies claimed they would. The piece focuses on Block’s decision to cut 40% of its workforce after projecting gross profits of nearly $12 billion for 2026. A technology leadership analyst quoted in the piece asks directly: “Who at that decision table was talking about the human cost of firing 4,000 people?” The piece concludes that the move was economically unnecessary given the company’s financial position. The question the piece raises is the one the Forrester research flagged earlier this year. Companies are making permanent workforce decisions based on AI efficiency projections that have not yet arrived. In other words, the AI tools that justified the cuts are still being assessed. The workers who were cut are gone. The Running Total Just Hit 142,000 Tech Times reported [https://www.techtimes.com/articles/317392/20260529/tech-layoffs-reach-142000-2026-profitable-companies-cut-jobs-fund-700b-ai-infrastructure.htm] on May 29 that tech sector layoffs in 2026 have reached 142,000, with Amazon, Microsoft, Alphabet, and Meta committed to a combined $700 billion in capital expenditure. Stanford HAI data cited in the piece shows software developer employment for workers under 26 fell nearly 20% since 2024. Wharton management professor Peter Cappelli describes the dynamic [https://www.techtimes.com/articles/317392/20260529/tech-layoffs-reach-142000-2026-profitable-companies-cut-jobs-fund-700b-ai-infrastructure.htm] as companies announcing layoffs by claiming AI will cover the work while the AI has not yet done so. The Workers Least Able to Recover Are the Ones Nobody Is Talking About A Brookings Institution and Centre for the Governance of AI analysis [https://www.brookings.edu/articles/measuring-us-workers-capacity-to-adapt-to-ai-driven-job-displacement/] published in February identifies 6.1 million American workers who face both high AI exposure and low adaptive capacity. These are secretaries, office clerks, payroll processors, receptionists, and tax preparers. Of these workers, 86% are women. Financial analysts scored 99% for adaptive capacity on the study’s measure. Office clerks scored 22%. These occupations are concentrated in smaller cities, university towns, and midsized markets in the Mountain West and Midwest. These workers are not the demographic the AI labor debate centers on. The debate centers on software engineers, lawyers, and accountants because those are the people writing about it. The 6.1 million clerical and administrative workers facing maximum exposure and minimum ability to recover are not writing Substack posts about their situation. They are answering phones, processing invoices, and preparing tax returns in Boise and Cheyenne and Lansing, and the policy response building in Sacramento and Washington is not designed with them specifically in mind. BUY NOW [https://wimdodson.gumroad.com/l/invasion_ai?_gl=1*g1iey2*_ga*OTc1NjU2NTcyLjE3NzQwMzA3NjM.*_ga_6LJN6D94N6*czE3NzQwMzMyMjUkbzIkZzEkdDE3NzQwMzQ2ODgkajYwJGwwJGgw]! Get the book that examines how the AI invasion already happened. You just weren’t invited. $9.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required. 2-hr reading time. They Are Already Using the Tools That Threaten Their Jobs The American Society of Administrative Professionals 2026 State of the Profession Report [https://www.asaporg.com/articles/the-ai-conversation-is-missing-something-about-administrative-professionals/] finds that more than three-quarters of administrative professionals already use AI daily in their work. It’s the same population that Brookings identifies as most exposed and least able to adapt. Ironically, it is also the population most actively integrating AI into daily tasks. They are using the tools. They are learning the workflows. They are doing what every piece of career advice tells them to do. The AI/Labor Report is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Using AI daily, however, does not close the gap between a 22% adaptability score and a 99% one. It does not replace the savings, the educational credentials, the broad skill set, or the access to a strong urban job market that the Brookings measure identifies as the actual determinants of who recovers when their role disappears. Get full access to The AI/Labor Report at ailabor.substack.com/subscribe [https://ailabor.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

IMF says AI is shrinking the middle class; 134,000 tech jobs gone this year; Upwork fires the workers it sells work to; Citi cutting 20,000; two CEOs soften their AI Jobpocalypse story before they IPO

The Middle Class Is the Target The International Monetary Fund does not typically frame its research in personal terms. Its January 2026 “Staff Discussion Note” on AI and new job creation [https://www.imf.org/en/publications/staff-discussion-notes/issues/2026/01/09/bridging-skill-gaps-for-the-future-new-jobs-creation-in-the-ai-age-572136] does. https://www.imf.org/en/publications/staff-discussion-notes/issues/2026/01/09/bridging-skill-gaps-for-the-future-new-jobs-creation-in-the-ai-age-572136The IMF finds that in regions with higher demand for AI-related skills, employment levels are 3.6% lower in occupations that are highly exposed to AI but offer limited scope for human-AI complementarity. The workers absorbing those losses are in entry-level and middle-skilled roles. Listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/the-ai-labor-report/id1896663061] About one in ten job vacancies in advanced economies now demands at least one new AI-related skill [https://www.imf.org/en/blogs/articles/2026/01/14/new-skills-and-ai-are-reshaping-the-future-of-work]. Those vacancies pay more. The jobs that do not require new AI skills are paying less and disappearing faster. So, AI skill premiums benefit workers at the top of the wage distribution. Workers at the bottom are also benefiting indirectly through service consumption. The group in between absorbs the displacement without capturing the gains. That group is the middle class. The IMF is an institution that chooses its words carefully. It used the phrase “contributing to the shrinking of the middle class” in a published research note. JPMorgan Just Put $40 Million on the Table for the Workers the IMF Is Describing Jamie Dimon arrived at a similar conclusion through a different route. In March, the JPMorgan CEO declared that the American Dream was “slipping out of reach for too many people.” This week JPMorgan published details of a $40 million small-business investment program [https://fortune.com/2026/05/27/jamie-dimon-american-dream-jpmorganchase-40-million-small-business/]. The program routs grants through community development financial institutions toward the communities his bank’s own data identifies as falling behind. JPMorgan’s own research finds that fewer than 10% of new businesses reach $1 million in revenue within five years. The structural disadvantage is concentrated among founders who lack inherited wealth or strong professional networks. The Freelance Economy’s Version of Work Degradation The online contract platform Upwork based its entire business on the premise that human knowledge work is abundant. It just cut a quarter of its own staff. Upwork CEO Hayden Brown announced on May 7 [https://officechai.com/ai/upwork-lays-off-25-of-its-workforce-says-ai-will-lead-to-smaller-teams/] that the company would reduce its workforce by approximately 25%. Her memo, published on Upwork’s website, stated: “Two pizza teams are dead. AI means smaller, differently resourced teams in product and engineering can make a bigger impact than ever.” BUY NOW! [https://wimdodson.gumroad.com/l/gods_in_the_machine?_gl=1*19da5a0*_ga*MTEwMjE2MDkwMy4xNzc2OTY4MDQy*_ga_6LJN6D94N6*czE3NzY5NjgwNDEkbzEkZzEkdDE3NzY5NjgxNzIkajU5JGwwJGgw] Get the NEW Book that exposes the Narratives Tech uses to build its AI Empire. $4.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required.3.5-hr reading time. The irony in the announcement is that Upwork exists to connect businesses with independent knowledge workers: writers, designers, developers, analysts, and marketers. Those are precisely the categories where AI has displaced freelancers most aggressively. Writing projects on Upwork fell 32% year over year in 2025. [https://www.selfemployed.com/news/ai-freelance-platforms-2026/] The CEO of the platform built on freelance demand is cutting staff because the demand for the work the platform sells has contracted. The displacement has an additional dimension. A customer support specialist or content writer displaced from one platform cannot simply apply for the same role at a competitor. Every competitor is reducing that function simultaneously. https://www.vaasblock.com/news/ai-layoffs-workforce-restructuring-cloudflare-coinbase-2026/Read my substack article “How to Ford ‘Amodei’s Moat:’ A Worker’s Guide to the AI Labor Shift [https://futureforwarded.substack.com/p/how-to-ford-amodeis-moat-a-workers]“ to find out how and why AI is making a career change in the labor marketplace has become more difficult. Nevertheless, the supply of mid-market writing, design, and support work is declining across the entire market, not redistributing to a different platform. [https://www.vaasblock.com/news/ai-layoffs-workforce-restructuring-cloudflare-coinbase-2026/] That is the structural unemployment dimension of 2026 that the aggregate statistics are not capturing. BUY NOW [https://wimdodson.gumroad.com/l/invasion_ai?_gl=1*g1iey2*_ga*OTc1NjU2NTcyLjE3NzQwMzA3NjM.*_ga_6LJN6D94N6*czE3NzQwMzMyMjUkbzIkZzEkdDE3NzQwMzQ2ODgkajYwJGwwJGgw]! Get the book that examines how the AI invasion already happened. You just weren’t invited. $9.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required. 2-hr reading time. The Running Total Just Crossed 134,000 A real-time tracker updated today [https://skillsyncer.com/layoffs-tracker]puts the 2026 tech sector layoff total at 134,603 workers across 212 layoff events since January 1. The pace is faster than the same period in 2025. Companies are simultaneously cutting in content creation, customer support, data entry, and basic coding while expanding in AI safety, machine learning operations, and AI-human collaboration roles. The workers being cut and the workers being hired are different people with different skills in different cities. The tracker number has grown by more than 20,000 since Tuesday. Citigroup Is Executing the Final Phase of Its 20,000-cuts Job Plan Citigroup filed WARN notices in New Jersey this week [https://www.thelayoff.com/citigroup] covering separations scheduled between May 21 and June 14, 2026, as part of an AI automation and restructuring strategy. [https://www.thelayoff.com/citigroup] The filings are part of Citi’s multi-year plan to eliminate 20,000 jobs across its global workforce by the end of this year. CFO Mark Mason has stated that headcount will keep declining as AI tools take hold across middle-office and operational functions. WARN notices are public legal documents. They put job losses on the record in a way that earnings call language does not require. The Citigroup filings this week describe tellers, back-office analysts, compliance staff, and IT support workers whose roles are being phased out as AI monitoring systems and automated processes replace them. These are the banking sector equivalents of the workers the IMF is tracking in its middle-class polarization data. The IPO Motive Behind the Week’s Altman Drama This week began with Sam Altman speaking to a group in Sydney and saying the jobs apocalypse he had warned about had not arrived. That story, covered in Tuesday’s edition, acquired a second dimension on Wednesday. Fortune reported [https://fortune.com/2026/05/26/sam-altman-dario-amodei-walking-back-ai-jobs-apocalypse-prophecies-ipo/] that both Altman and Anthropic CEO Dario Amodei have publicly reversed their most alarming predictions about AI job losses, with the timing coinciding directly with IPO preparations. OpenAI is targeting a late 2026 public listing at a valuation near $1 trillion. Anthropic is planning a 2026 offering at approximately $380 billion. Amodei previously warned that 50% of white-collar jobs faced elimination within several years. He now frames automation as a productivity multiplier. “If you automate 90% of the job,” he said this month, “then everyone does the 10% of the job.” The 10%, he argues, expands to fill 100% of what people do and multiplies their output tenfold. The argument is coherent as economic theory. It describes a future in which workers remain employed and become more productive. It doesn’t seem, though, that many CEOs are treading the “pro-worker” path of AI use. The Yale Budget Lab Data Altman and Amodei Are Leaning On The aggregate stability data that both CEOs are now citing as evidence their earlier predictions were wrong comes from the Yale Budget Lab’s ongoing labor market tracker. The Yale research [https://mlq.ai/news/altman-and-amodei-walk-back-ai-job-apocalypse-warnings-ahead-of-trillion-dollar-ipos/]finds no significant shifts in occupational mix or unemployment for high-AI-exposure jobs since ChatGPT’s 2022 launch. The AI/Labor Report is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. The same research notes that occupational shifts already visible in 2021 were underway before generative AI tools became widely available. The Yale data measures employed workers in formal occupations. It counts people who are in the system. It does not count the contractors, gig workers, and the freelancers who used to work on Upwork. Further, it does not account for the early-exit buyout takers who have already left formal employment and fall outside the tracking algorithms the aggregate statistics rely on. The stability finding is accurate for the population the Yale Report measures.However, that population is getting smaller. Get full access to The AI/Labor Report at ailabor.substack.com/subscribe [https://ailabor.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

Microsoft's AI chief puts an 18-month countdown on your job; AI wins 83% of real professional tasks; Goldman counts 16,000 jobs lost monthly; South Korea plans for job disruptions, D.C. doesn't care.

The Man Who Runs Microsoft’s AI Division Just Put a Countdown on Your Job On Tuesday, Sam Altman spoke to executives in Sydney and said the jobs apocalypse was overblown. On Wednesday, the man who runs Microsoft’s AI division contradicted him directly. Mustafa Suleyman told the Financial Times [https://finance.yahoo.com/sectors/technology/articles/microsoft-ai-ceo-prediction-future-154055391.html] that most white-collar professional tasks will be fully automated by AI within 12 to 18 months. He named the roles specifically: lawyers, accountants, project managers, marketing professionals. His exact words were “human-level performance on most, if not all, professional tasks.” Suleyman is the CEO of Microsoft AI, a division of the company that simultaneously offers buyouts to 8,750 employees. He knows what the company’s AI systems currently do. His 12-to-18-month window is the most specific timeline any sitting AI executive has attached publicly to mass white-collar automation. Listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/the-ai-labor-report/id1896663061] The contrast with Altman’s Tuesday remarks is direct and consequential. Altman said he was “delighted to be wrong” about the speed of job losses. Suleyman said the timeline for displacement is now measured in months. Both men are building the same technology. They are describing two different futures for the same workforce. Challenger, Gray & Christmas confirmed this week [https://afrotech.com/microsoft-ai-ceo-predicts-ai-replace-white-collar-tasks] that AI was the leading stated reason for corporate job cuts for the second consecutive month. AI-attributed layoffs reached 49,135 in 2026 so far. The firm’s chief revenue officer Andy Challenger put the core mechanism plainly: “Regardless of whether individual jobs are being replaced by AI, the money for those roles is.” What Goldman Sachs Found When It Actually Counted Goldman Sachs Research published an analysis this month [https://www.goldmansachs.com/insights/articles/the-jobs-ai-is-likely-to-boost-and-those-it-may-disrupt] reported that roughly 16,000 net jobs per month over the past year. The methodology separates AI substitution (where AI replaces workers entirely) from AI augmentation (where AI assists workers and can increase demand for human labor). Substitution is winning. The net effect raised the U.S. unemployment rate by 0.1 percentage points. Goldman economist Joseph Briggs stated that if AI job losses arrive faster than the bank’s base-case projection [https://ai2.work/blog/goldman-sachs-says-ai-job-losses-could-force-the-fed-to-cut-rates], the labor market deterioration could become severe enough to force the Federal Reserve to cut interest rates.BUY NOW! [https://wimdodson.gumroad.com/l/gods_in_the_machine?_gl=1*19da5a0*_ga*MTEwMjE2MDkwMy4xNzc2OTY4MDQy*_ga_6LJN6D94N6*czE3NzY5NjgwNDEkbzEkZzEkdDE3NzY5NjgxNzIkajU5JGwwJGgw] Get the NEW Book that exposes the Narratives Tech uses to build its AI Empire. $4.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required.3.5-hr reading time. So, in roles where AI assists rather than replaces workers, employment and wages are rising. But the augmentation gains are concentrated in a smaller set of occupations than the substitution losses. The labor market is sorting into winners and losers faster than the aggregate statistics reveal. The Benchmark That Made Suleyman’s Claim Concrete Carnegie Endowment for International Peace cited a new OpenAI economist study [https://carnegieendowment.org/research/2026/04/the-ai-labor-debate-three-views-on-the-future-of-work] this month finding that current AI models outperformed human workers on 83% of 220 high-value professional tasks. The tasks were selected from the 44 occupations responsible for the largest share of U.S. GDP. They averaged seven hours to complete. They were written and graded by professionals with an average of 14 years of industry experience. AI won or tied on 83% of the tasks. The benchmark covers legal analysis, financial modeling, project scoping, and technical writing. It covers the specific job categories Suleyman named. Altman’s claim that the apocalypse is smaller than he feared reflects the aggregate unemployment rate. The OpenAI benchmark reflects what AI systems can do today on the actual tasks those workers perform. BUY NOW [https://wimdodson.gumroad.com/l/invasion_ai?_gl=1*g1iey2*_ga*OTc1NjU2NTcyLjE3NzQwMzA3NjM.*_ga_6LJN6D94N6*czE3NzQwMzMyMjUkbzIkZzEkdDE3NzQwMzQ2ODgkajYwJGwwJGgw]! Get the book that examines how the AI invasion already happened. You just weren’t invited. $9.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required. 2-hr reading time. South Korea Is Building a National Response. California Has a 180-Day Study. A Carnegie Endowment report published this month [https://carnegieendowment.org/research/2026/04/from-labor-scarcity-to-ai-society-governing-productivity-in-east-asia] documents the policy gap between East Asia and the United States. South Korea’s 2026 National AI Action Plan tasks five separate government ministries with building retraining hubs, vocational conversion programs, an AI Employment Service Roadmap, and compensation plans for workers displaced by AI. China’s Ministry of Human Resources announced a forthcoming national document on AI’s labor market impact with programs to stabilize employment. Meanwhile, Japan’s government committed to addressing labor shortages through coordinated AI and robotics deployment. The AI/Labor Report is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. California’s response, covered in this space on Wednesday, is a 180-day study order. South Korea’s response is a multi-ministry deployment with compensation mechanisms already under design. The two approaches describe governments with fundamentally different assessments of how much time workers have before organizations replacing staff with AI. Oracle’s decision to cut between 20,000 and 30,000 workers globally [https://thenextweb.com/news/oracle-layoffs-march-2026] while posting a 95% jump in net income highlights the international policy gap. Approximately 12,000 of those cuts landed in India, Oracle’s largest offshore engineering hub, where a national AI labor protection framework does not yet exist. Oracle co-CEO Mike Sicilia stated publicly that AI coding tools now enable “smaller engineering teams to deliver more complete solutions more quickly.” Who, then, will be left to maintain, update, and interact with human customers? AI Agents? Get full access to The AI/Labor Report at ailabor.substack.com/subscribe [https://ailabor.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]

Class of 2026 hits decade-high unemployment; PayPal cuts 4,760 for AI savings; Orgs pay AI layoff survivors more and employs fewer; California takes 180 days to study obvious AI worker disruptions

The Degree That Was Supposed to Protect You For most of the past three decades, the standard advice to American workers was direct: get a college degree, move into knowledge work, and stay ahead of the automation wave. That advice produced a generation of workers who did exactly what they were told. The Federal Reserve Bank of New York now reports [https://www.newyorkfed.org/research/college-labor-market] that the unemployment rate for recent college graduates reached 5.6% in March, one of the highest levels in a decade outside of the pandemic. The national unemployment rate stands at 4.3%. New graduates are now faring worse than the general population, a reversal of a pattern that held for thirty years. The Class of 2026 walked across their stages this month into a market that looks nothing like the one their advisors described when they enrolled four years ago. One Worker Who Lost His Job This Week Andrew Tran is 40 years old. He was a product designer at Meta. He lost his job this week [https://www.cbsnews.com/news/ai-layoffs-hiring-entry-level-workers/], part of the 8,000-person cut the company executed after redirecting its labor budget toward AI infrastructure. Tran told CBS News he plans to find work at a company using AI “intentionally,” rather than as a tool for replacing workers. He said corporations “should have an obligation to retrain their workforces instead of throwing them to the curb.” Tran represents the demographic the data has been pointing at all year. He is educated, experienced, and employed in the kind of knowledge-work role that was supposed to be safe. He is also 40, which puts him in the age bracket Anthropic’s own labor market research [https://www.anthropic.com/research/labor-market-impacts] identifies as among the most AI-exposed. The workers the technology hits hardest are more likely to be female, over 40, more educated, and better paid than the workers people assumed would be first in line. Listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/the-ai-labor-report/id1896663061] PayPal Plans to Cut One in Five Workers Over the Next Three Years Yahoo Finance [https://finance.yahoo.com/markets/stocks/articles/paypal-layoffs-ceo-cuts-20-154944985.html] reported that PayPal plans to cut roughly 4,760 workers over the next two to three years. That’s about 20% of its current staff level. PayPal processes payments for millions of American small businesses and independent contractors. The company’s new leadership frames the reduction as part of an AI-driven simplification of operations. Forrester identified earlier this year as a strategy for wage arbitrage dressed up as efficiency [https://www.forrester.com/report/workforce-ai-displacement-2026/]. That slow-burn layoff pattern is the version of displacement most likely to disappear inside the aggregate statistics. BUY NOW! [https://wimdodson.gumroad.com/l/gods_in_the_machine?_gl=1*19da5a0*_ga*MTEwMjE2MDkwMy4xNzc2OTY4MDQy*_ga_6LJN6D94N6*czE3NzY5NjgwNDEkbzEkZzEkdDE3NzY5NjgxNzIkajU5JGwwJGgw] Get the NEW Book that exposes the Narratives Tech uses to build its AI Empire. $4.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required.3.5-hr reading time. California Just Became the First State to Order a Formal Response to AI Worker Displacement Governor Newsom signed Executive Order N-6-26 on May 21 [https://www.gov.ca.gov/2026/05/21/governor-newsom-signs-first-of-its-kind-executive-order-to-prepare-workers-and-businesses-for-potential-ai-disruption/]. The order directs California state agencies to develop policies, gather data, and identify early warning signs of AI-driven workforce disruption. BUY NOW [https://wimdodson.gumroad.com/l/invasion_ai?_gl=1*g1iey2*_ga*OTc1NjU2NTcyLjE3NzQwMzA3NjM.*_ga_6LJN6D94N6*czE3NzQwMzMyMjUkbzIkZzEkdDE3NzQwMzQ2ODgkajYwJGwwJGgw]! Get the book that examines how the AI invasion already happened. You just weren’t invited. $9.95 flat fee for Kindle, Nook, Tablets, and Mobile. No subscription required. 2-hr reading time. It calls for recommendations within 180 days on updates to California’s WARN (Worker Adjustment and Retraining Notification) Act. WARN currently requires employers to notify workers of large layoffs 60 days in advance. The order also directs agencies to study severance standards, employment insurance, transition support, and worker ownership models. California Labor Federation’s President Lorena Gonzalez said the order is welcome but added that “it’s not enough to just study the issue, we have to take action now.” California is the state with the most AI companies, the most AI workers, and now the most visible political pressure on an executive who wants to run for president in 2028. AI and worker protections looks to be sizing up (along with data centers) to be a main plank of the Democratic Party in the run up to the 2028 Presidential election. California’s Legislature Passed the No Robo Bosses Act — The Same Bill Newsom Already Vetoed Two days before Newsom signed his executive order, the California Senate passed SB 947, the No Robo Bosses Act, 29-9 [https://sd05.senate.ca.gov/news/ca-senate-approves-no-robo-bosses-act-2026-ensure-human-oversight-ai-workplace]. The bill bars employers from relying solely on AI automated decision systems to fire or discipline workers and requires human oversight in termination decisions. The bill now moves to the state Assembly. Newsom vetoed a nearly identical version of this bill last fall. He cited concerns about overly broad restrictions on employers. The current version was revised to address those objections. The question of whether he signs or vetoes it a second time will determine whether the largest state economy produces any binding protection against automated termination decisions before the EU’s equivalent rules take effect in August. The AI/Labor Report is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Federal Reserve Data Shows What “Stable” Actually Means Federal Reserve Bank of Dallas research [https://www.dallasfed.org/research/economics/2026/0224] published this year finds that employment in AI-exposed sectors trails the broader economy while wages in those same sectors grow faster than the national average. Since fall 2022, national nominal wages rose 7.5%. Wages in the computer systems design sector rose 16.7% over the same period. So, AI is suppressing the number of workers employed while concentrating wage gains among those who remain. Fewer people share a larger total wage pool. The aggregate wage statistics look healthy. The employment statistics in AI-exposed sectors look different. Federal Reserve Bank of New York economists confirmed in May [https://libertystreeteconomics.newyorkfed.org/2026/05/do-job-postings-show-early-labor-market-effects-of-ai/] that the effect is already visible in job posting data. Vacancy patterns between AI-exposed and non-exposed occupations are diverging. The signal appears in hiring before it appears in unemployment data. That means the official unemployment statistics are trailing the actual disruption by an unknown lag. The Same Pattern, 9,000 Miles Away A UPI report published Monday [https://www.upi.com/Top_News/World-News/2026/05/25/gig-economy-worker/5161779758231/] warns that approximately 40 million gig economy workers across Southeast Asia face AI-driven automation exposure with no meaningful social safety net in place. A McKinsey survey found two-thirds of major Southeast Asian companies have already fully adopted AI or are actively expanding its use. A speaker at a semiconductor conference in Kuala Lumpur last week said the automation trend “has become irreversible.” The 40 million figure describes workers doing the same kinds of tasks that American contractors, platform workers, and call center employees perform. They exist outside any formal employment relationship. They fall outside displacement tracking systems. They face no WARN Act, no severance study, and no executive order. They’re on their own. Get full access to The AI/Labor Report at ailabor.substack.com/subscribe [https://ailabor.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_4]