Asking About AI and the Fall of the Middle Class
Today I want to discuss the development of artificial intelligence and the decline of the middle class. When will that decline begin in earnest, what will it do to society, and how should we prepare? This is a day for questions, but first we need to think through this larger current together.
This is not speculation about a distant future. It is already happening before us, and the change is moving faster than politicians, educators, and economic experts can follow.
The fall of the middle class will not end with a few successful people losing their jobs. It will shake the labor market, consumption, money and ownership, education, and even personal identity.
A YouTube Thumbnail in Minutes
Rather than merely explain the point, I used ChatGPT image generation to make the thumbnail for this broadcast. After I entered the topic and wording, AI planned the composition and placed an anxious face, disappearing jobs, the title, and supporting text in a single pass.
In the past I would have spent one or two hours finding images, cutting and combining them in Photoshop or Krita, adjusting colors and backgrounds, and arranging type. That also required design judgment and mastery of the tools. Now an almost finished result appears in minutes.
All I did was state my request as though I were speaking to an expert. That is why this is not merely a change in technology but a change in the structure of work.
Toad, Take My Old House and Give Me a New One
A Korean children’s rhyme says, “Toad, toad, take my old house and give me a new one.” We have long imagined a magical being that hears a wish and returns a result. AI now plays the role of that toad: I hand over an old request and it returns a new artifact.
Sufficiently advanced science is hard to distinguish from magic. In twenty or thirty years, fantasy-like acts—throwing fireballs, moving objects, or seeing through a bird’s eyes—may become real through nanomachines, AI, and advanced devices.
The problem is that an age in which everyone becomes a magician is not necessarily a happy one. At the same time, the experts who once performed the magic lose their value.
Experts Were Society’s Toads
Lawyers, judges, doctors, accountants, professors, programmers, designers, musicians, actors, directors, writers, and architects were society’s toads. They converted human requests into results through knowledge and skill acquired over years of study and failure.
A doctor narrows symptoms and test results into possible diagnoses. A lawyer searches statutes and precedents, and a programmer turns requirements into code. Much of human intellectual labor has consisted of prediction and judgment based on accumulated patterns.
AI learns those patterns more broadly and retrieves them faster. That is why intellectual labor contained within a screen is displaced before robots can replace physical labor completely.
The Entire Apprenticeship Is Skipped
Translation, drawing, editing, composing, and coding once required learning tools and failing repeatedly until the work became second nature. That long apprenticeship separated experts from everyone else. Now people can describe a desired result and let AI perform the intermediate process.
A Korean speaker can question foreign texts and receive help as though a personal lawyer, doctor, professor, and translator were always present. Knowledge is changing from the property of a limited group into something like a universally accessible public good.
As differences in knowledge and skill shrink, the foundation on which the middle class maintained higher income and status also weakens. The democratization of technology expands opportunity while dismantling established expert groups.
Creative Work Is Not Exempt
People will be able to create the martial-arts film they want to see, the novel they want to read, or the music they want to hear through dialogue with AI. A film once required dozens or hundreds of actors and crew, but even broadcasters now fill reenactments and reference footage with AI-generated material.
This does not simply mean that every creator vanishes. It means that far fewer people are needed to produce the same volume of work and that the entry-level rungs disappear first.
Jobs for reenactment actors, camera operators, photographers, editors, writers, and composers decline together. A magical creative tool empowers one person while removing many positions from the labor market.
Leveled Knowledge and Meaningless Exams
AI is already approaching or surpassing most humans on standardized, college-entrance-style questions. Soon it will instantly provide knowledge broader and deeper than that of teachers, professors, and many domain experts.
What is the point of ranking people by memorization and correct answers in such an age? If human performance is no longer needed, grading people from first to last by that performance also loses its purpose.
The important ability will not be the quantity of stored facts but knowing what to ask, how to inspect an answer, and how to connect different fields. Critical thought and the ability to question must move to the center of education.
Even My Programming Skill Is Losing Its Use
I spent ten or twenty years learning to build software. That ability supported my work for a long time, yet the share of code written and repaired by AI is rising rapidly. The skill I spent a lifetime building is not exempt.
English is similar. I invested enormous time in reading English when I was young, and it helped with programming documents, news, and games. Since translation quality improved in my late forties and early fifties, however, I have rarely read English documents in the original.
I now read far more philosophy, education, religion, history, and even Latin texts. It is not because I learned Latin but because AI translates them. I use the old skill less while reaching more knowledge—a central paradox of this change.
Education Is Still Stuck in the Twentieth Century
Parents repeat the path they know: teach English and mathematics, raise scores, enter a prestigious university, and secure a good job. Yet we first need to ask whether that job will exist when the child graduates.
The number of students and teaching positions is shrinking, but young people still spend four years training at teachers’ colleges and then plead for hiring. The system keeps consuming their lives in preparation for disappearing positions.
Schools lack both the confidence to redesign the system and the capacity to match the pace of change. Some education that develops thought remains valuable, but most memorization aimed at entrance scores and rankings will lose its use.
Ask Why Universities Exist, Not How to Make Ten Elite Ones
A policy such as “make ten Seoul National Universities” might have helped twenty years ago. Today it copies the old university at the very moment the need for universities themselves is in question.
Universities once drew power from gathering good professors, libraries, and information in one place. When a personal AI can connect world literature and lectures and explain them individually, the concept of the university must change completely.
Counting and ranking institutions without answering why they should exist is obsolete nonsense. Universities may disappear while policy is still counting how many elite ones to make.
At Most Three Years Until the Shock Becomes Obvious
Others say five or ten years, but at the present rate I believe the decline of the middle class could become unmistakable within three years at most, perhaps two. From April 2026, that means roughly 2028 to 2029.
Not every occupation will vanish on a single day. Markets for junior lawyers, accountants, developers, translators, and creators are already narrowing. Entry-level jobs disappear first, cutting the ladder through which the next generation gains experience.
Presidents, legislators, scientists, educators, and economists cannot follow the pace. The end of labor is approaching, yet they repeat old stories instead of sounding the alarm.
We Do Not Need to Wait for Physical AI
It is far too late to assume that disaster begins only after robots occupy every factory and street. High-income knowledge workers can be displaced by software before embodied robots arrive.
Much of the work of doctors, lawyers, accountants, developers, and designers is information processed inside computers. That is why junior professionals cannot find jobs and developers are dismissed even though physical AI is not yet widespread.
Physical AI must be manufactured and tested and faces battery and safety constraints, so it will spread more slowly than software. The decline of the middle class will not wait. It begins with white-collar displacement and radiates outward.
Blue-Collar Workers Are Not Safe
The collapse of white-collar work does not leave blue-collar work untouched. Displaced professionals must survive, so some enter cleaning, transport, and factory labor. When more people compete for fewer positions, wages and conditions deteriorate.
That is the balloon effect: labor squeezed out of one area presses on another. At the same time, large factories adopt robots first, while autonomous vehicles threaten the livelihoods of huge numbers of drivers.
When I researched the subject, roughly a million people in Korea made a living from driving. Vehicle automation alone can produce an enormous shock. We cannot comfort ourselves that blue-collar displacement is more than a decade away.
Mass Unemployment Beyond the IMF Crisis
The full decline of the middle class could bring an unprecedented period of chaos and pain. The shock may dwarf what Korea experienced in the 1997 financial crisis: bankrupt companies, forced retirement, a surge in self-employment, and the spread of dual-income households.
In that crisis, labor was still needed once the economy recovered. This time the issue is not lost productive capacity but a structural reduction in the need for human work.
If people are not needed, they also lose the route by which they earn income. Left alone, mass unemployment becomes more than personal tragedy; it leads to the breakdown of the economic system itself.
When the Middle Class Falls, Consumption Stops
Who buys cars, films, travel, coffee, cosmetics, medicine, and leisure? The middle class stands at the center of consumption beyond bare survival. Poor people must spend nearly all they have simply to live.
If the middle class loses income, it does not matter how cheaply corporations produce through robots and AI; nobody can buy the goods. Without consumers, money stops circulating, and production and investment stop as well.
An age of mass unemployment passes through collapsing consumption into a crisis of the economic and monetary systems. That is why we cannot watch the decline of the middle class as someone else’s misfortune.
Basic Income Becomes a Tool for Preserving the System
When money stops circulating, economic and monetary systems fail, and an angry public may ask why a tiny group owns everything. To prevent revolt and keep consumption alive, society will be forced to discuss basic income.
Basic income is not exclusively a left-wing demand. Even highly right-wing AI CEOs and wealthy owners may support higher taxes and basic income. If the economy and the state collapse, the law that protects their property becomes a scrap of paper.
For the wealthy, basic income may be less charity than insurance for property and markets. It distributes enough money for people to eat and consume so that the system survives.
Can We Accept Permanent Class Division?
Those who own resources, land, energy, and AI can use robots to produce almost anything. Those without ownership have no way to earn money when their labor is no longer required.
If money does not circulate from the top downward, there is no mechanism for closing the gap. Society may harden into a minority that owns everything and a majority living on basic income.
Will people forever accept a system that says, “We live in heaven and own everything; you take the allowance we give you”? Basic income may be necessary, but it cannot eliminate the vast conflict over permanent class division.
More Than Occupations Will Collapse
Someone may study for a lifetime to become a lawyer, only to discover upon qualifying that few lawyers are needed. People who devoted themselves to becoming doctors, painters, composers, or programmers may face the same fate.
An occupation supplied more than income. It formed self-respect, identity, and a plan for the future. A person who struggled into the middle class may be unable to bear suddenly becoming no different from everyone else.
Relative deprivation and despair may drive some people to suicide. Some indicators might eventually improve if everyone were similarly poor, but the transitional path to that condition is the greatest danger.
White-Collar Decline Degrades the Whole Labor Market
A translator is an advanced knowledge worker who must understand Korean, a foreign language, and context. Yet AI now handles most basic translation, leaving newcomers with no entry point.
Development is similar. When fewer people train as developers, coding academies close. When a profession contracts, the education, equipment, and service ecosystem around it contracts too.
When displaced white-collar workers compete for cleaning and factory jobs, conditions for existing blue-collar workers worsen. This chain reaction has already begun and cannot be reduced to “a few professionals losing work.”
The Labor Movement Must Change Its Goal
A labor movement that only demands better treatment for workers cannot follow an age in which labor itself disappears. The distinction between those who can work and those who cannot is breaking down.
The necessary movement must demand humane treatment not only for workers but also for people who do not work. Survival and dignity cannot be denied merely because no job is available.
Politicians, intellectuals, and educators are nearly silent, and the labor movement also misses the point when it repeats its old line. It must begin from the possibility that labor itself will vanish.
What Should We Teach a Child?
My child is still in third grade, so for now I mainly encourage reading rather than specialized training. Around fourth or fifth grade, and more seriously in middle school, I plan to develop habits of questioning and reflection.
I also try to persuade my wife that English academies, convoluted calculation problems, homework, and rankings do not deserve excessive weight. We cannot even know whether universities will exist in their present form ten years from now.
Schools may remain valuable less as the sole source of knowledge than as places to meet peers and converse. Communication and common life matter, but score competition has less and less reason to remain the center of education.
Learning Together With AI
With AI and computers, a person can move through conversations, maps, and documents in a day or two and study an amount that once required two or three books. We can handle an incomparable volume of material in a short time.
Teachers are unprepared because they were never educated in this way. Giving students a tool is not enough; they must learn to verify answers and connect contexts.
The central questions are what interests a learner, what to ask, and how critically to inspect an AI response. Thought will not simply stop. The methods of learning and education will change.
There Is No Frictionless Liberation
Once AI becomes social infrastructure, humanity may be freed from working like slaves merely to survive. A future of abundant production and widely available knowledge is genuinely attractive.
But immense friction lies on the path to it. No society has prepared for change this abrupt, while institutions and assumptions remain rooted in an older age.
We must see both the long-term liberation offered by technology and the pain of transition. Talking about a good future does not erase unemployment, inequality, or the collapse of identity.
A World That Changes Whenever We Open Our Eyes
Recent image, video, and music systems produce results that are hard to believe were made by AI. People once said the landscape changes in ten years; now it changes in a week.
We therefore have to understand the present in order to anticipate the future. If parents repeat old formulas for success and governments design policy around old universities and labor markets, the change itself will destroy people.
Toad, take my old house and give me a new one. The age in which everyone becomes a magician has already begun. Our task is not merely to admire and use the magic, but to perceive in advance the structures of life it will tear down.
This article was edited by AI solely from the supplied video captions and description. No external sources were used.