On the threshold of great mutations, there are always prophets of collapse. They announce the end of work, the end of history, the end of ideologies, the end of growth, the end of man. Today, more prosaically, they are announcing the end of work under the combined effect of artificial intelligence, physical robots and digital agents. Tomorrow, they say, machines will produce everything; whole swathes of the economy, even the highly valued digital economy, will fall into nothingness. Humans, now useless, will be relegated to forced idleness and misery, the nuisances of a robot society.
This vision, popularized recently by a number of alarmist reports, including most recently a note from Citrini Research, a small independent financial and macroeconomic analysis firm founded by an American financial analyst, Al?p Shah, provoked a stock market storm in mid-February 20206 by proposing a worst-case scenario in which AI causes billions of jobs to disappear as early as 2028, without creating as many. This diagnosis is based in particular on the launch this week of Anthropique’s Claude Code and Claude Cowork agents, capable of automating code, marketing, legal and data analysis tasks, destroying the entire software economy and causing immediate panic on the stock market.
We need to keep our wits about us. Artificial intelligence is a historical gas pedal. And like all gas pedals, it liberates as much as it destroys, it enables as much as it eliminates, it displaces far more than it destroys. And that’s what we need to understand.
Today’s anxiety has accompanied every previous major technological shift: when agriculture became mechanized, mass unemployment and the ruin of the peasants was predicted; they became workers. When industry automated, it was predicted that workers would starve; they became employees. When information technology invaded offices, it was predicted that employees would be ruined; they became service sector executives. Each time, the prophecy proved false: one world disappeared. Another was born.
Today, we’re witnessing the accelerating demise of repetitive, measurable, standardizable work – work that can be codified and optimized. Artificial intelligence (whether in the form of a robotic arm, a decision-making algorithm or a conversational agent) is attacking this particular form of work, and the doomsday scenarios that ensue are based on a simple illusion: that the jobs destroyed will not be replaced.
But this is never how the real economy works. Each wave of technical progress simultaneously produces three inseparable effects: it destroys existing jobs that are tedious or boring; it abruptly increases overall productivity; and it frees up colossal resources: capital, time, human energy and financial margins.
What the alarmist reports forget is that wealth does not disappear with the jobs it replaces. It changes form, ownership and destination. And, above all, it always seeks to be deployed elsewhere.
The Citrini Report repeats an old false prophecy: that of capitalism destroying itself for lack of consumers. In reality, when oligopolistic rents disappear, when intermediaries become unnecessary, when costs collapse, overall activity increases. The losers are visible; the winners are diffuse. The former shout; the latter consume, invest and invent.
Artificial intelligence does not eliminate economic dynamics. It shifts it, it recomposes it, it accelerates it. What’s more, the accelerated ageing of Europe, East Asia and, soon, China, is transforming robotization into a vital necessity: in these societies, AI does not replace existing workers. It prevents the collapse of essential services. It maintains living standards.
It stabilizes public finances. And, paradoxically, it allows humans to be re-employed where they are needed most: with other humans.
Fear feeds on the disappearance of jobs. Foresight then poses a revolutionary question: what to do with the money freed up in this way?
Every time productivity increases in sectors that can be automated, demand shifts to those that cannot. Yesterday, agricultural mechanization financed industry. Yesterday, industrialization financed hospitals and schools. Tomorrow, artificial intelligence will make it possible to robotize a large part of human activity, freeing up the funds to finance what I call the life economy.
This includes: sustainable energies, organic farming, healthy food, relational care, from infancy to the end of life; education, transmission, research, creation, rebellion, accompaniment of pathways; social, cultural and territorial mediation; symbolic, artistic and spiritual creation; democracy, ethical arbitration, responsibility, security, freedom, critical thinking, and everything that is useful for future generations.
And there is no limit to the desire to live longer, in better health, to eat better, to learn more, with more relationships, more meaning, more beauty, more physical and emotional security, more knowledge, more culture, more freedom.
These activities have four characteristics in common: they are intensive in terms of human presence; they are not very compressible in terms of productivity; they are structurally underfunded everywhere; finally, the machine, robot or AI, will never be able to accomplish these missions, not for technical reasons, but for anthropological ones: an airplane is not a bird. A robot will never be a human.
Artificial intelligence does not destroy the economy of life, which it cannot achieve. It just makes them financeable at last.
The danger is not that AI will destroy too many jobs. The danger is that companies will refuse to redistribute the profits that this job destruction will generate. If the resources freed up are captured by a minority, if taxation remains fixed on human labor when value migrates to algorithmic capital, if social recognition remains exclusively attached to commodity-producing market employment, then a tiny minority will be very rich, while a large majority will not have the means to access the goods of the life economy and will have a pitiful life.
This will not be the fault of AI and robots. It will be the fault of institutions and political choices.
So the real question is not: “Will AI take away jobs? But “what will we do with the abundance it makes possible?”
Image : AdobeStock