AI Won’t Take All Jobs, But Will Wipe out the Middle Class
AI utopia will split society: super-workers vs. stipend class
Elon Musk predicts that in less than 20 years, work could become optional as machines take over almost every task. Is this workless future really possible — and could it create more problems than it solves?
Musk’s prediction that robots will do all the work isn’t really a forecast; it’s a thought experiment. What’s far more likely is AI eating the middle. Admin, clerical, back-office, junior, and mid-level knowledge work — anything that boils down to “move information from one box to another” — we simply won’t need as many humans for anymore.
Are there things robots won’t do? Plenty. And there are things they might technically do, but we won’t let them. Can we build a robot that can change your baby’s diaper? Sure. But are you really going to hand your newborn to a steel arm with a firmware update? No, you want a human in the room. According to ELVTR’s new survey, over 80% of people say healthcare workers and judges must stay human-only. Around three-quarters feel the same about teachers, therapists, presidents, and executive branch leaders. In other words, if a job involves life, death, justice, or children, people want a human in charge. When things go wrong, people want a neck to wring, not a Terms of Service. I don’t see us accepting a world where our courtroom is run by a server rack. And we’ll keep paying for human presence in performance, sport, food, and culture; nobody is lining up yet to watch “Arsenal vs Algorithm 3.0.”
We will also always need plumbers and dentists — messy, embodied, in-person work that’s either too expensive to automate or that many of us simply won’t want to hand to machines.
Then, at the top, you get a narrow elite of super-workers who treat AI as an extension of their brain, not as a substitute. They orchestrate AI to build products, companies, and policy.
So what happens to the people whose jobs get eaten in the middle? Do we get a “stipend class” of economically useless citizens? That’s not science fiction; it’s an active discussion among economists: a future where a significant share of the population simply cannot sell their labor at wages that support a decent life, given how markets value tasks. In that world, society can absolutely afford to respond with some form of basic income, dividend, or stipend funded by taxes on extreme productivity and AI-generated wealth. The real conflict isn’t economic, it’s political and cultural — especially in the US. Can a country built on “you eat what you kill” live with tens of millions of citizens whose main economic role is to consume and not riot?
What do these people do with their lives? Some will be fine: a sort of permanent, low-grade retirement — streaming, gaming, caregiving, local projects, a long tail of “play work” and side hustles that are more about identity than survival. For some people, that’s heaven. For many, it’s corrosive. Humans aren’t built for pure consumption; we need to feel needed. If we don’t redesign education and institutions around meaning, agency, and the ability to work with AI, not just around content and credentials, we won’t just have a labor-market problem — we’ll have a quiet epidemic of boredom, resentment, and addiction.
Written by Roman Peskin, CEO of ELVTR — a leading platform for AI and next-level professional skills — warns that this “robot utopia” could quietly become a social catastrophe