Everyone has a relative who predicted something, got ignored, and then spent the next decade being very right and very gracious about it. Technology tends to produce a version of that person at scale. When the first waves of serious AI automation started rolling through offices and factories and call centers, the loudest skeptics were often older workers, people who had watched previous technological overhauls up close and had opinions about the gap between what the brochure promised and what actually happened to the people standing in the way of the efficiency gains. Those opinions were not always welcome. The word “alarmist” got used a lot.
The pitch, as it has been for every major technological shift since at least the industrial revolution, was overwhelmingly optimistic. AI would create more jobs than it destroyed. Workers would be freed from drudgery and elevated to more meaningful work. Productivity gains would lift wages across the board. These were not cynical claims; many of the people making them genuinely believed them. But there is a difference between a forecast and a plan, and as the AI workplace impact has become impossible to ignore in 2026, the distance between those two things has become very clear.
The skeptics had a list. Not a formal one, but a consistent set of worries that kept surfacing in op-eds and retirement dinners and HR town halls and conversations that got shut down before they could really get going. Here is how the list has held up.
1. Jobs Would Disappear – and Not Always Be Replaced by Something Equal
The argument was always that automation creates as many jobs as it destroys. That is technically still the forecast. The World Economic Forum’s Future of Jobs Report 2025, which drew on surveys of over 1,000 employers representing more than 14 million workers worldwide, projected that 92 million roles will be displaced by 2030 while 170 million new roles emerge. A net gain of 78 million, on paper, sounds reassuring.
The part that does not make the headline is the gap between those two events. The roles being displaced exist right now, and the people in them are not automatically the people who will fill the new ones. A data entry clerk displaced in 2025 does not become an AI systems trainer by next quarter. The transition has a cost that falls on real people, and the optimistic long-term math does not absorb that cost. About 1 in 6 employers already expect AI to reduce headcount in 2026, and reports indicate AI contributed to 4.5% of total job losses in 2025.
The Boomers who watched the auto industry’s gutting of entire rust-belt communities after the first wave of industrial automation were not being paranoid when they said the replacement jobs tend to be somewhere else, in a different sector, requiring different skills, available to different people.
2. White-Collar Workers Would Not Be Immune
The early narrative about automation was that it would come for manual labor first and knowledge work last, if at all. The legal secretary worried. The attorney was told not to. That framing did not hold.
In the first six months of 2025 alone, 77,999 tech job losses were directly attributed to AI. White-collar workers in financial services and media now express higher levels of concern about automation – 67% – than workers in transportation (60%) or retail (59%). The people who were told their cognitive work was safe have spent the last two years watching AI draft contracts, generate financial reports, write code, and synthesize research at speeds that no individual human can match.
This is exactly what several generations of skeptics warned about when they pushed back on the “AI will only replace routine tasks” reassurance. Routine turns out to be in the eye of the beholder, and the beholder, increasingly, is a very fast and cheap algorithm.
3. Surveillance Would Become the Default, Not the Exception
The monitoring started reasonably enough. Remote work during the pandemic meant companies wanted visibility into whether work was actually getting done, and that seemed defensible. The tools that arrived to enable that visibility did not go away when offices reopened. They expanded.
A February 2025 survey of 1,500 employers and 1,500 employees found that 74% of US employers now use online tracking tools, including real-time screen tracking by 59% and web browsing logs by 62%. These are not passive tools. AI-powered monitoring software can log keystrokes, flag email content for sentiment, track how long a worker is away from their desk, and generate productivity scores that feed directly into performance reviews. The employees subject to this often have no meaningful way to contest the score.
61% of Americans oppose employers using AI to track workers’ movements, and 56% oppose AI monitoring desk presence. The gap between what employees want and what employers are deploying is enormous, and it is not narrowing. The people who said that giving companies digital visibility into workers’ every minute would fundamentally change the nature of the employment relationship were not wrong.
4. The Wage Gap Between the Tech-Savvy and Everyone Else Would Widen
The promise was rising tides. Better tools for everyone, productivity gains shared across the workforce, higher wages flowing from efficiency. The reality has been more uneven. A 2026 Springer Nature review of the research literature from 2020 to 2025 found that AI adoption is closely associated with rising demand for technical and interdisciplinary skills and with widening wage gaps between AI-skilled and non-AI-skilled workers.
Workers who can build, deploy, or fluently work alongside AI systems command a significant premium. Workers whose jobs are being partially or fully automated face wage stagnation at best and displacement at worst. The research confirms that AI technologies increase salary compensation for higher skills, widening the gap between workers with different skill levels, and that middle-skilled workers, whose tasks are often the easiest to automate, are most at risk.
This is the version of “AI creates new jobs” that the fine print describes. The new jobs pay extremely well. There just aren’t that many of them, the path to getting one is not obvious, and the people losing their old jobs are not automatically first in line.
5. Hiring Would Become Less Human – and That Would Create New Problems
Before AI entered recruitment at scale, hiring was already imperfect and biased. But it was imperfect in ways that could be challenged, examined, and argued with. A human interviewer could be questioned. An algorithmic rejection is a closed door with no address on it.
A large-scale randomized experiment published by VoxDev found that leading AI models systematically favor female candidates while disadvantaging Black male applicants, even when qualifications are identical. The study, examining five leading large language models, concluded that intersectionality, rigorous piloting, and human oversight are all necessary to reduce algorithmic bias in hiring.
The AI hiring tools in use today were trained on historical employment data. That data reflects decades of human bias. The AI learned from it. The result is a system that can launder discrimination through an algorithm, making it harder to name, harder to challenge, and harder to correct. The people who said “what happens when no one is accountable for the decision” were asking the right question.
6. Entry-Level Positions Would Dry Up – Hurting Young Workers Most
Entry-level jobs have always been imperfect. The pay was low, the work was often tedious, and the prestige was minimal. But they served a function that no amount of training programs has yet replicated: they were the place where people without experience could get experience. AI is filling many of those roles now, and the ladder that generations of workers climbed is missing some rungs.
A 2025 Stanford study reported by CBS News tracked the impact of generative AI on early-career employment and found that workers between the ages of 22 and 25 in AI-exposed fields like customer service, accounting, and software development saw a 13% decline in employment since 2022, while employment for more experienced workers in the same fields held steady or grew. In software engineering and customer service specifically, entry-level employment fell roughly 20% between late 2022 and mid-2025.
The result is a generation of young workers who cannot build the same kind of career foundation that their predecessors took for granted. The Boomers who warned that automating the bottom rungs would strand young people had a point – they just could not have predicted it would be knowledge work, not factory work, where the problem would hit hardest.
7. Entire Industries Would Restructure – With or Without a Plan
Restructuring sounds clinical. What it means, at the human level, is that the organization chart your department worked under last year may look entirely different this year, and nobody paused to ask the people doing the work how the transition should go. AI adoption has accelerated the pace of organizational change faster than most companies’ HR functions have been able to respond.
The World Economic Forum found that 41% of employers globally plan to reduce their workforce in areas where AI can automate tasks within the next five years. That is not a fringe plan – it is the majority position among large employers. What it means for the workers in those areas is rarely communicated clearly in advance, and the reskilling programs that are supposed to absorb the displaced are chronically underfunded and inconsistently delivered.
The people who said “companies will adopt the technology first and figure out the human consequences later” were describing, with some precision, what is currently happening.
8. Creative and Knowledge Work Would Lose Its Protections
For most of the automation conversation, creative professionals, writers, designers, researchers, and educators were considered largely safe. The argument was that genuine creativity and original thought could not be replicated by a machine. That boundary has been, at minimum, significantly complicated.
Generative AI now writes marketing copy, produces visual assets, drafts legal briefs, generates lesson plans, and synthesizes research papers. Whether it does any of these things as well as a skilled human is a conversation worth having. But it does all of them cheaply and fast, and “as good as a junior professional” has become the threshold at which employers start making staffing decisions. The volume of work that previously required a team can now be handled by one person with the right tools and two hours.
The Boomers who predicted that no job was truly safe from automation were considered alarmist. The evidence from 2024 and 2025 across the creative industries suggests they were simply early.
9. Worker Burnout Would Increase, Not Decrease
The pitch was that AI would take the repetitive drudgery off workers’ plates, leaving them free for more rewarding, higher-level tasks. The actual experience for many workers has been that AI tools require constant management, the remaining work has become more cognitively dense, the pace has accelerated because efficiency gains are immediately converted into higher output expectations, and the headcount has often been reduced so that fewer people are doing more.
Springer Nature’s review found that many jobs are not eliminated but transformed, as AI frequently functions in augmentation rather than pure substitution mode. Augmentation sounds positive, and often is. But it also means that the worker is now expected to manage the output of AI tools, quality-check them, correct their errors, and explain them to clients – while also doing the work itself, at the same or faster pace than before, with the same or fewer colleagues.
The pressure created by AI-augmented work is real, and it is not distributed equally. Workers with less institutional power have less ability to push back when efficiency expectations outpace human capacity.
10. The Skills Landscape Would Change Faster Than Education Could Track
Schools, universities, and vocational programs are slow. They take years to redesign curricula, get new programs accredited, hire qualified instructors, and graduate enough students to move the needle on the workforce. AI is not slow. The gap between the skills the economy needs and the skills the education system is producing has widened substantially in a very short period.
The IMF found that the demand and supply of new skills, especially in IT and AI, are reshaping labor markets and affecting wages and hiring – and that about one in ten job vacancies in advanced economies now demands at least one new skill, often appearing first in the United States. Those vacancies go unfilled not because the workers don’t exist but because the training pipelines haven’t caught up.
The people who said “we’re going to need to rethink education from the ground up, and that takes time we might not have” were not being dramatic. The speed of adoption has outpaced every institutional mechanism designed to help workers adapt.
11. Accountability Would Get Blurry
Before AI made consequential decisions, there was at least a human somewhere in the chain who could be identified as having made a call. A manager who passed you over for a promotion. A loan officer who denied your application. The systems weren’t always fair, but there was someone to look at. AI decision-making has substantially complicated that picture.
When an algorithm rates an employee’s performance, denies a loan application, screens out a job candidate, or flags a worker for termination review, the question of who is responsible for that decision becomes genuinely difficult to answer. Was it the algorithm? The company that licensed it? The vendor that built it? The training data that shaped it? States are beginning to catch up – states like Colorado and New York are now requiring employers to disclose and audit AI tools used in employment decisions – but the legislative response is running behind the deployment.
The Boomers who said “if a machine makes a decision about my livelihood, I want a person I can talk to” were articulating something important about due process and dignity that many organizations are still figuring out.
12. The Benefits Would Not Reach Everyone Equally
This was perhaps the most consistent concern across generations of technology skeptics: that the gains from new technology tend to concentrate at the top. The productivity improvements go to shareholders and executives. The displacement falls on workers. The people who benefit most from AI are the people who already had the most – access to education, capital, and institutional power.
Research shows that 79% of employed US women work in high-automation-risk jobs, compared to 58% of men. The administrative, clerical, and customer service roles that AI is automating most aggressively are disproportionately held by women, and those workers have the least institutional support for navigating the transition. The technology is not neutral. It arrives on top of an existing structure of advantage and disadvantage, and it amplifies what it finds there.
The distribution of AI’s benefits and burdens follows lines that were entirely predictable and that were, in fact, predicted. That those predictions came from people who were told they didn’t understand technology, and that those people turned out to be right, is something the industry has been quietly reckoning with ever since.
The Record Speaks for Itself
None of this is an argument against technology. It is not a case for going back to typewriters or pretending that automation hasn’t produced genuine improvements in medicine, safety, research, and quality of life. The argument is simpler than that: the concerns were real, they were specific, and the people raising them deserved to be taken more seriously than they were.
What the last several years of accelerating AI deployment have demonstrated is that the optimistic framing – “it will create more than it destroys,” “no one who wants to adapt will be left behind,” “the efficiency gains will benefit everyone” – was incomplete. Not entirely wrong, but incomplete in ways that have real consequences for real people. The fact that 170 million new jobs are projected to exist by 2030 is cold comfort if you are 52 years old, worked in accounts receivable for twenty years, and the company’s AI implementation eliminated your department in Q3 of last year.
The lesson is not that skeptics are always right. It’s that when a technology is being deployed at this scale and this speed, the people asking hard questions about who bears the cost of the transition deserve something better than reassurance. They deserved a plan. In most places, that plan is still being written. And the people who said so first are watching, without particular satisfaction, from the sidelines.
AI Disclaimer: This article was created with the assistance of AI tools and reviewed by a human editor.