The AI/Labor Report — Friday, 12 June 2026

The AI/Labor Report — Friday, 12 June 2026

Customer service is the largest single employment category in the American economy. It is also the one AI is targeting most directly, most publicly, and soonest. Forrester’s 2026 CX Outsourcing Report [https://www.cmswire.com/customer-experience/ai-will-eliminate-nearly-half-of-all-customer-service-jobs-by-2030/] projects that 49% of current customer service jobs will be gone by 2030. High-volume consumer contact centers face 80% AI containment within five years. Lower-volume B2B environments face roughly 70% containment because their cases require more judgment and less repetition. If AI handles eight or nine out of every ten contacts, the workforce required to handle the remainder looks very different from what exists today. Listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/the-ai-labor-report/id1896663061] Verizon illustrates how that process unfolds on the ground. Before CEO Dan Schulman announced the largest round of layoffs in Verizon’s history [https://www.thestreet.com/employment/verizon-ceo-cuts-to-the-chase-new-layoffs-ai-future], the company required customer service staff to participate in mandatory AI role-playing exercises. Sales representatives spent weeks interacting with an AI system the company described as a tool for improving customer service skills. A former employee reported that the sessions began six weeks before Schulman announced 13,000 job cuts [https://www.wirefly.com/news/report-verizon-employees-may-have-trained-ai-thats-replacing-them] and grew more intensive as the layoff date approached. Verizon said the training was not designed to gather replacement data. The timing made that statement difficult to accept at face value. The company has since confirmed a second round of cuts [https://memeburn.com/verizon-ai-will-replace-much-of-customer-service-in-2026/] scheduled for August 7 at its New Jersey headquarters, with no retraining programs announced for affected workers. Gartner has identified the pattern [https://www.metaintro.com/blog/ai-not-replacing-jobs-2026-data-gartner] clearly enough to project its next phase. A survey of 321 customer service leaders found that only 20% of companies actually reduced staffing because of AI, with most 2025 layoffs tied to federal policy changes and post-pandemic right-sizing instead. Gartner projects that half the companies that did cut workers for AI will rehire them by 2027, but under new job titles. The returning workers will manage and coach the AI systems that replaced their former colleagues. The job is the same conversation, at one remove, with a fraction of the original headcount. Gartner also projects [https://www.retellai.com/blog/will-ai-replace-call-center-agents] that conversational AI will reduce contact center labor costs by $80 billion in 2026, even though only one in ten agent interactions will actually be automated this year. The savings come not primarily from firing staff but from AI absorbing repetitive, high-volume contacts that currently drive 30 to 45% annual staff turnover across the industry. Companies do not need to announce layoffs to shrink. They simply stop replacing people who leave. 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 workers best positioned to survive this shift are those who have built genuine fluency with AI tools. Microsoft’s 2026 Work Trend Index [https://smart-team.io/en/work-trend-index-2026-microsoft/], surveying 20,000 knowledge workers across 10 countries, found that only 19% of AI users operate in what Microsoft calls the Frontier Zone, where individual capability and organizational support reinforce each other. Ten percent are in a state of blocked agency: workers with genuine AI skills trapped inside organizations that have not updated their processes or incentives to put those skills to use. Only 13% say their organization explicitly rewards reinventing work with AI. Meanwhile, Google announced Thursday a $50 million commitment [https://www.axios.com/2026/06/11/google-trade-worker-initiative-ai] to train more than 300,000 skilled trade workers across the U.S., directing funds through 14 labor unions and four trade associations. The investment targets the electricians, fiber technicians, welders, and HVAC specialists needed to build and maintain AI data centers. Meta announced a $115 million training program for skilled trade workers earlier this week. Anthropic committed $150 million to a national fellowship program. An estimated 2.1 million skilled trades jobs could go unfilled nationally by 2030. The Trump administration’s immigration policies have accelerated the shortage, with construction jobs bearing the heaviest impact of any sector. The companies building the infrastructure that replaces service workers cannot find enough people to build the buildings that house it. 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 most authoritative counterargument to the Forrester projections comes from the OECD’s June 2026 Economic Outlook [https://fortune.com/2026/06/03/oecd-economic-outlook-scarring-effects-recession-ai-labour-displacement/], titled “Under Pressure.” The report found no signs of widespread labor displacement from AI at the industry level. Job vacancies in AI-exposed sectors grew faster than in other sectors. The primary constraint, the OECD argues, is a shortage of workers who can use AI rather than a surplus of workers it has replaced. Unemployment among workers aged 20 to 30 in AI-exposed roles, though, rose approximately 3 percentage points in 2025. 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 aggregate numbers look stable because experienced workers are absorbing more output per person. The entry door for young workers trying to build careers in AI-exposed fields is closing. The OECD finds no widespread displacement. What it finds instead is a generation trying to get their first foothold in a labor market that is quietly pulling the ladder up. 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]

Amazon's new robot speaks and responds to orders; India's entry jobs collapsed 44% yoy; Japan embraces robots as it depopulates; South Africa hits 45% youth unemployment Artificial Immigrants arrive

On June 4, Amazon gathered journalists at a fulfillment center near London to introduce the next generation of its Proteus warehouse robot [https://www.techtimes.com/articles/318088/20260609/amazon-unveils-conversational-proteus-warehouse-robot-10b-europe-push.htm]. The new model can roam freely across an entire facility and take instructions in plain conversational language. Workers tell it what it needs to do. It figures out the route, the timing, and the priority on its own. Amazon also announced a plan to invest more than €10 billion modernizing its European fulfillment network and add 25,000 jobs across the continent. At roughly the same time, Amazon CEO Andy Jassy sent a memo to staff [https://www.cnbc.com/2026/06/05/amazon-robot-proteus-warehouse-ai-layoffs.html] acknowledging that AI will reduce the company’s total corporate workforce over the coming years. “We will need fewer people doing some of the jobs that are being done today,” he wrote. Listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/the-ai-labor-report/id1896663061] Amazon has eliminated tens of thousands of corporate positions since late 2025. The London event was a showcase for the physical side of that same process. Robots that understand plain language do not need specialized software operators. The work of directing them moves from a trained technician to anyone on the floor, and then, eventually, to no one in particular. The workforce consequences of this pattern are already visible in India, where Amazon cut 12,000 jobs in April as it redirected spending toward AI infrastructure. India’s technology job market hit a 28-month low in June [https://analyticsindiamag.com/ai-news/india-tech-hiring-hits-28-month-low-as-active-job-demand-falls-17-yoy/], with only 93,000 active openings recorded, down 14% from May and 17% from a year earlier. IT services, the sector that built India’s modern middle class, saw openings fall 31% year over year. Entry-level positions for workers with up to two years of experience dropped to 10,000 openings [https://www.outlookbusiness.com/corporate/india-incs-fresher-hiring-sees-steep-drop-amid-ai-adoption], a 44% annual decline. EY India’s technology sector leader summed up the shift plainly: revenue growth is now decoupled from headcount growth. Indian IT firms are expanding output without expanding staff. The same dynamic runs through Southeast Asia, where it lands on workers with far fewer protections. A McKinsey survey [https://www.upi.com/Top_News/World-News/2026/05/25/gig-economy-worker/5161779758231/] found that about two-thirds of major Southeast Asian companies have already fully adopted AI or are actively expanding its deployment. At the Semicon Southeast Asia 2026 conference in Kuala Lumpur, industry leaders described the automation trend as permanent. “There is no going back,” one speaker said. An estimated 40 million gig workers across the region face direct exposure to these shifts. A World Bank report urged governments to expand AI job training and strengthen support for displaced workers, while noting that many countries in the region still lack basic AI skills education in their school systems. Japan faces the same technology at the opposite end of the demographic spectrum. Japan has an estimated 1.3 million unfilled technology positions and a labor shortage [https://siliconcanals.com/sc-w-japan-is-deploying-robots-not-to-replace-workers-but-because-theres-no-one-left-to-hire/] so severe that companies are deploying robots primarily because there are not enough humans available to hire. A Reuters/Nikkei survey confirmed that labor shortages, not cost-cutting, are the primary force pushing Japanese firms toward AI and robotics adoption. Toyota, Mitsubishi Electric, and Honda are scaling physical AI systems as a workforce substitute rather than a workforce reduction. Japan is the only major economy in which AI deployment functions as a solution to a labor shortage rather than a cause of one. In Africa, the problem runs in yet another direction. At the 2026 Future of Jobs Summit [https://news.sap.com/africa/2026/05/2026-future-of-jobs-summit-why-sas-future-workforce-must-adapt-faster-to-survive-the-ai-revolution/]in Sandton, South African business leaders and policymakers described an urgent collision between a worsening unemployment crisis and accelerating AI adoption. Youth unemployment among South Africans aged 15 to 34 stands at 45.8%. The jobs that AI is beginning to reach in South Africa never fully materialized for young workers there in the first place. A report published in Ghana this week [https://www.modernghana.com/news/1498226/africa-is-not-losing-jobs-to-ai-it-is-losing.amp] argues that Africa’s real AI labor risk is not mass displacement but a failure to build skills fast enough. Youth unemployment among Ghanaians aged 15 to 35 already stands near 21%. 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 World Economic Forum’s 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/] projects that automation will displace 92 million jobs globally by 2030 while creating 170 million new roles. Those new roles, however, require advanced digital proficiency that most African workers currently have no path to acquire. The UN Economic Commission for Africa [https://www.uneca.org/stories/ai-and-jobs-in-africa-turning-disruption-into-opportunity] convened more than 200 participants earlier this year to examine whether AI will drive job creation or labor market disruption across the continent. The participants agreed that the benefits will not arrive automatically or reach workers evenly. In Europe, a new layer of worker-facing regulation took effect this week with almost no fanfare. The European Commission published a Code of Practice [https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai] on marking and labelling AI-generated content on June 10. The Code directly affects content workers, journalists, and media professionals operating in EU markets. The week’s stories span four continents and describe four distinct versions of the same pressure: too many robots arriving too fast in some places, not enough workers available in others, and almost nowhere a legal structure prepared to manage either outcome. 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. 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]

The AI/Labor Report — Wednesday, 10 June 2026

GitLab [https://www.bbntimes.com/technology/gitlab-cuts-14-of-workforce-to-fuel-ai-agentic-era-ambitions] reported revenue of $264 million last quarter, up 23% year over year. It beat analyst expectations. Then it laid off 350 people. The company calls its restructuring GitLab Act 2 [https://about.gitlab.com/blog/gitlab-act-2/], a repositioning for what it describes as the “agentic era” of software development, in which autonomous AI agents handle coding, reviewing, deploying, and maintaining software under human oversight. The savings from cutting 350 jobs go directly into GitLab’s AI agent platform, Duo, which generated roughly $20 million in committed recurring revenue after its first full quarter. Management expects no material revenue from Duo in fiscal 2026. A profitable company is trading its current human workforce for a product that will not pay off this year. Listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/the-ai-labor-report/id1896663061] That pattern repeats across the industry with reliable consistency. The organization does not need to be struggling. Revenue growth is beside the point. What drives the restructuring is the calculation that AI agents will multiply productivity per remaining employee while reducing the total headcount required. GitLab is simply one of the most candid companies about naming that logic publicly. The wage data coming out of Europe and the United Kingdom shows the downstream consequence for individual workers. An Oxford University study of more than 10 million UK job postings [https://www.weforum.org/stories/2026/02/ai-improving-wages-job-quality/] found that candidates with AI-related skills command a 23% wage premium on average. A Master’s degree carries a 13% premium. A Bachelor’s degree carries 8%. AI proficiency now outperforms formal educational credentials in immediate labor market returns. The same research found that AI skills helped older workers and candidates without advanced degrees improve their hiring prospects substantially when those skills appeared on a resume, particularly when supported by a recognized certificate. The catch is in what happens to the workers on the other side of that divide. An IMF staff discussion note published this year [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] finds that about one in ten job vacancies in advanced economies now demands at least one new AI-related skill. Those vacancies post higher wages. They also correlate with lower employment in occupations with high AI exposure and low complementarity with AI tools, posing direct challenges for younger workers entering those fields. The labor market is not simply rewarding AI fluency. It is simultaneously defunding the entry points that workers once used to build the experience that eventually earns those higher wages. A LinkedIn report presented at Chatham House on June 9 [https://www.chathamhouse.org/events/all/standard-event/ai-work-and-future-global-competitiveness] puts a headline number on the productive upside: AI could unlock up to $6.6 trillion in capacity for businesses in the U.S., UK, France, Germany, and India if implemented across all work tasks. The LinkedIn framing is the one corporate communications departments will circulate internally. The IMF framing is the one workers in AI-exposed roles should read alongside it. Both documents describe the same economy. They assign the gains and the costs to different people. Today, GitLab holds its Transcend conference [https://www.thestreet.com/employment/github-rival-shuts-offices-cuts-100s-of-workers-gitlab], where the company is expected to release the next wave of its agentic AI product roadmap. The announcements will name specific developer tasks that AI agents are slated to absorb next. That list will be worth tracking closely, because each item on it represents work that currently requires a human being. The political response to all of this arrived from an unexpected direction this week. Senator Bernie Sanders announced the American AI Sovereign Wealth Fund Act [https://fortune.com/2026/06/03/bernie-sanders-ai-ownership-sovereign-wealth-fund-electrification/], which would impose a one-time 50% tax on AI company stock paid in shares, depositing that equity into a public fund giving ordinary Americans voting rights, board representation, and eventually a direct payment. The proposal is built on a straightforward argument: AI is built on publicly funded research, trained on publicly generated data, and deployed in ways that displace workers who have no recourse under current law. The public, Sanders argues, should own a share of what it helped create. The AI/Labor Report is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. OpenAI CEO Sam Altman met privately with Sanders [https://www.bostonglobe.com/2026/06/06/nation/trump-sanders-and-altman-all-talking-about-public-ownership-in-ai/] and said he also wants the public to have equity in AI companies, though he declined to support the 50% threshold. President Trump told reporters Friday [https://fortune.com/2026/06/05/trump-partnership-openai-anthropic-xai-nationalization-bernie-sanders-altman/] that the U.S. government may take direct equity stakes in OpenAI, Anthropic, and xAI, calling it “a beautiful thing.” The left and the right and the CEO of the most powerful AI company in the world are now circling the same populist anxiety. None of them have agreed on anything. Workers displaced by the tools those companies sell are still waiting for a law that requires anyone to tell them why. 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. 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]

Uber fires HR Division; Uber's AI budget gone in four months; Manila's 2M jobs at risk; OpenAI Codex plugins target job titles; California slow-walks for AI/jobs protections

Uber had a busy week. On June 3, the company cut 23% of its People and Places division [https://www.cnbc.com/2026/06/03/uber-layoffs-people-division-ai.html]. The division covers human resources, recruiting, workplace facilities, and corporate culture. The official explanation was organizational complexity and overlapping responsibilities. The company was careful to say artificial intelligence played no role in the decision. Days earlier, Uber’s chief technology officer disclosed the company had burned through its entire 2026 AI coding budget in four months [https://www.techtimes.com/articles/317842/20260605/uber-cuts-23-hr-staff-new-president-denies-ai-role-95-engineers-use-it-daily.htm]. Ninety-five percent of Uber’s engineers use AI tools every month. Close to 70% of committed code is now generated by AI systems. Uber still lists more than 800 open positions. The company is reorganizing, not shrinking. The division being restructured happens to be the one that manages human workers. Now, AI tools handle a large share of the work human workers used to support. Earlier in the year, Deutsche Bank analysts coined a term for this pattern: AI redundancy washing [https://www.ibtimes.co.uk/uber-ai-budget-job-cuts-1801340]. It describes companies that use AI as the stated reason for cuts when the real drivers are more complicated. The Uber case is the cleanest example yet of both dynamics running at the same time. On June 2, OpenAI announced that Codex, its AI coding agent, now has more than 5 million weekly users [https://openai.com/index/codex-for-knowledge-work/]. Developers are still the largest group, but knowledge workers now represent 20% of users and are growing three times faster than developers. To accelerate that shift, OpenAI released six role-specific plugins targeting data analytics, creative production, sales, product design, equity investing, and investment banking [https://techcrunch.com/2026/06/02/openai-launches-new-codex-tools-for-white-collar-work/]. Each plugin integrates directly with coporate tools like Salesforce, Jira, and Notion. The move mirrors a similar enterprise push from Anthropic earlier this year. So, OpenAI has designed tools that approximate specific jobs. The list of roles in those six plugins covers a substantial portion of what mid-level office workers at large companies spend their days doing. One in five Codex users today is already a non-developer. That share is growing faster than any other segment. Governments in general are watching the AI drama unfold and are still moving slowly. California is moving faster than most. On May 21, Governor Gavin Newsom signed Executive Order N-6-26 [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/], directing state agencies to evaluate whether the state can update California’s WARN Act to catch AI-driven workforce reductions before they happen rather than after. WARN is the law that requires advance notice for mass layoffs Recommendations to the governor’s office are due in 180 days. A separate California bill, SB 951 [https://www.reedsmith.com/our-insights/blogs/employment-law-watch/102mx9t/whats-new-in-californias-response-to-ai-workforce-disruption/], would require 90 days advance notice before any AI-driven layoff affecting 25 or more workers. The federal government has no equivalent law. No employer anywhere in the United States currently faces a legal obligation to disclose whether AI drove a workforce reduction. In Europe, the regulatory picture shifted in the opposite direction. In May, the European Parliament and Council reached a provisional agreement to delay the EU AI Act’s employment enforcement deadline [https://www.fisherphillips.com/en/insights/insights/eu-overhauls-ai-act-just-before-key-deadline] by 16 months, moving it from August 2, 2026 to December 2, 2027. The rules being delayed cover AI systems used in hiring, performance evaluation, worker monitoring, promotion, and termination decisions. Employers using those systems in Europe now have more time to comply. Workers in those same workplaces have 16 more months without the protections those rules would have provided. The international dimension of this story reaches further than European regulatory calendars. A Harvard Business Review analysis published June 5 [https://www.365outsource.com/virtual-assistance/ai-powered-assistants-philippine-outsourcing/] argues that generative AI is dismantling the labor-cost arbitrage that built the global outsourcing industry. 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 Philippines employs 2 million people in business process outsourcing and generates $40 billion annually from it. India’s sector employs 6 million and contributes 7% of GDP. Industry analysts project 2 to 3 million workers across both countries face disruption by 2030 [https://news.outsourceaccelerator.com/ai-threats-bpo-india-philippines/]. That means roughly 1 million BPO and IT jobs at direct risk in India. Oracle already cut 12,000 jobs in India in April as it redirected spending toward AI infrastructure. The layoff mechanism operating in Manila and Bengaluru is the same one operating in Uber’s People and Places division. The work that once required a human in a lower-cost location now requires a software subscription. The geography of the cost savings has changed, but the math facing workers has not. 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]

Grads earn $56K not $80K; 42.5% of grads underemployed; remote work pulled up job ladder, AI chipping away at higher rungs; AI agents fail 19 of 20 tasks; FOMO begetting FOBO

Graduation season arrived this year with an unusual tradition. Commencement speakers who mention artificial intelligence keep getting booed. The students are not being dramatic. The unemployment rate for recent college graduates hit 5.6% in the first quarter of 2026 [https://www.cnbc.com/select/class-of-2026-hiring-stats-and-ai-trends/], the highest level in a decade outside the pandemic spike. The national unemployment rate is 4.3%. Recent graduates are now doing worse in the job market than the average American worker, which has not been the normal pattern for people with four-year degrees. The underemployment rate for new graduates reached 42.5%, the highest since 2020. Listen on Apple Podcasts [https://podcasts.apple.com/us/podcast/the-ai-labor-report/id1896663061] Graduates expect to earn $80,004 in their first year out of college. Actual starting salaries average $56,153 [https://www.cnbc.com/select/class-of-2026-hiring-stats-and-ai-trends/]. Applications per job posting jumped 26% while early-career postings fell 16%. The Class of 2026 is walking into a labor market that is hiring less, paying less, and competing more than the one their older siblings entered. Remote Work Created the Problem. AI Is About to Make It Worse. The Federal Reserve Bank of New York published a study on June 1 [https://fortune.com/2026/06/01/new-fed-study-remote-work-ai-driving-higher-unemployment/] attributing 64% of the rise in young graduate unemployment to remote work rather than AI. Companies are reluctant to hire inexperienced workers into roles where mentoring and on-the-job training are difficult to deliver remotely. Unemployment among graduates under 29 rose 20% after the pandemic. Unemployment among graduates over 29 in the same fields fell slightly over the same period. The Fed’s claim is that new workers need proximity to learn, and remote work removed that proximity. The New York Fed researchers noted their own finding may be temporary. Stanford researchers found that early-career workers aged 22 to 25 in the most AI-exposed jobs saw employment fall 16% since late 2022 even after stripping out remote-friendly roles. Nevertheless, the aphorism that arose during the peak-Covid remote work experience applies here to new job entrants: If you can do you can do your work remotely, so can a machine. Jobs Are Not Disappearing. They Are Shrinking. A May 2026 Gartner report covered by The Register [https://www.theregister.com/ai-and-ml/2026/05/06/ai-layoffs-backfire-as-cutting-staff-doesnt-cut-it-firms-warned/5230631] describes AI’s effect on employment as hollowing out rather than elimination. AI absorbs discrete tasks within existing roles, narrows the scope of work, and compresses wages without necessarily removing the job title from the org chart. Workers whose roles depend on judgment, context, and accountability may find AI a useful tool. Workers whose roles depend on processing, routing, summarizing, and formatting are finding themselves doing less of that work, earning wages that reflect the reduced scope, and uncertain about how the change happened. Imperial College London and Microsoft jointly found [https://www.theregister.com/ai-and-ml/2026/05/06/ai-layoffs-backfire-as-cutting-staff-doesnt-cut-it-firms-warned/5230631] that AI adoption paradoxically increases workplace burdens for workers who remain, as their jobs shift toward reviewing, verifying, and correcting the output of multiple AI agents. Gartner forecasts that new job categories will emerge from this transition by 2028 or 2029. That is a gap of two to three years during which the hollowing out continues without the offset. When Every Company Automates at Once, Everyone Loses A paper published June 3 by University of Pennsylvania researchers [https://arxiv.org/abs/2603.20617] took at at AI FOMO (Fear of Missing Out) by herds of corporate execs as potentially harmful to their own companies interests. Each individual company rationally reduces labor costs through AI adoption without waiting to see if the AI will actually do the jobs. When all companies do it simultaneously, consumer spending contracts, revenue bases erode, and the productivity gains from automation are collectively undermined. Each company individually believes it is making the rational balance-sheet decision to lay off staff. Bureau of Economic Analysis data cited in the paper [https://arxiv.org/abs/2603.20617] shows business investment overtook consumer spending as the primary driver of U.S. GDP growth in Q1 2026. The personal savings rate fell to 3.6% in March, its lowest since late 2022. Workers are spending down savings while companies invest upward. The productivity gains are flowing into capital formation. They are not reaching household income. The Agents Companies Are Replacing Workers With Cannot Do the Work A CBC News investigation published June 3 [https://www.cbc.ca/news/world/ai-agents-tech-company-layoffs-9.7221069] pulled together the controlled performance evidence on AI agents. Carnegie Mellon’s TheAgentCompany benchmark tested 10 leading AI agent models across 175 real-world office tasks. The best-performing model completed 30.3% of tasks successfully. Gartner found AI agents produce incorrect results on office tasks roughly 70% of the time. A major AI infrastructure company stated its agents fail to produce professionally acceptable work more than 19 times out of 20. Companies are replacing workers with these tools right now. Block’s stock jumped 20% when it announced AI-attributed layoffs. Future Forwarded is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. A University of Toronto researcher told CBC that the explanation is straightforward: “Investment in AI as an ideology has more value than keeping your workforce. That’s how capitalism works.” Workers who remain after AI-related cuts find their jobs have changed. Their new role is reviewing, correcting, and monitoring the agents that replaced their former colleagues. 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. 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]