§Essay

Artificial Intelligence and Work

Rahim Taghizadegan · 5 min read

The debate about artificial intelligence and work is usually conducted at the wrong level. People ask which jobs will disappear, which new ones will emerge and which skills will be in demand in future. These questions are reassuring, because they suggest that this is just another round of a historical game that modern societies know well: technological upheavals destroy some activities, create others, and in the end a new equilibrium sets in.

The question that actually matters is a different one: what happens to a society when a large part of its members can, for structural reasons, no longer make any positive contribution to the creation of value? Not temporarily, not for lack of training, but permanently.

Seen historically, this worry seems exaggerated. The decline of agricultural employment long looked like a dead end. Illiterate people used to physical labour seemed hardly suited to industrial or administrative work. And yet the transformation succeeded. Similar fears accompanied every great technological break. In hindsight, each transition appears almost inevitable.

What is often overlooked is the decisive difference from the present. In earlier upheavals it was possible either to lower the cognitive demands of productive work or to raise the cognitive level of broad sections of the population. Institutions, incentives and tools really did contribute to this. The so-called Flynn effect was no myth. Societies could count on productivity and cognitive ability rising together.

That option is now gone. With AI agents, for the first time a factor of production appears that is not merely faster or cheaper, but cognitively superior in principle. And not in a few specialist fields, but across almost all intellectual work that can be formalised. The hope that humans will find new roles at the interfaces rests on a picture of limited sensors, expensive robots and a physical world that is hard to model. These remaining domains are now coming under pressure as well. The focus thus shifts from retraining to marginal productivity. Societies have always lived with people whose integration into productive processes required more coordination, supervision and training than it created in value. The institutional expression of this is well known: sheltered employment, charitable support, workshops for the disabled. A combination of care, discipline and symbolic production. What is new is not their existence but their possible scale.

Already today, a considerable part of the population in the West lives not on productivity that the market will pay for, but on redistribution. And this redistribution is not confined to social transfers. Money creation, credit expansion, subsidies, regulatory privileges and state-financed consumption distort demand and employment deep into the supposedly private economy. Many white-collar jobs exist not because someone pays voluntarily for their output, but because institutional flows of payments sustain them. "Bullshit jobs" is less a moral accusation than a description of missing validation by the market.

Artificial intelligence will destroy little here. Where activities are cartelised, rising productivity leads not to job cuts but to denser procedures, more regulation and new reporting duties. If the effort required of civil servants, lawyers or compliance departments falls, their positions do not disappear. They expand.

The real break therefore lies not in the loss of work as such, but in the loss of a central social illusion. The modern order rests on the assumption that in principle everyone can be integrated productively, provided incentives and education are right. Artificial intelligence undermines this idea not out of malice but through superiority. It makes visible what was long concealed.

As bleak as this diagnosis sounds, it is itself based on a further, rarely questioned assumption: that a society is stable only if as many people as possible are held in formal employment. This idea is historically young. It is a product of the modern central state. Making paid employment universal served human flourishing less than it served the creation of a stable tax base. Work was to be visible, measurable and tied to a place. It is no accident that a job is also called a "position".

Earlier societies dealt differently with a surplus of people. Surplus children vanished into monasteries, onto pilgrimages, onto trading voyages or into the discovery of new worlds. We no longer have surplus children anyway.

In this light, artificial intelligence appears less as a threat than as the last realistic hope. Without massive gains in productivity, the demographic catastrophe will crush civilisation. An ageing, shrinking population can only be relieved by tools that replace real work. Add to this another contradiction: the taxation of the family forces both parents into paid work, even where this makes no economic sense. Often enough the job of the one, by way of redistribution, pays for the job of the other, instead of balance and care being organised more efficiently within the family.

Less work would therefore be no catastrophe but excellent news. At least for families. The precondition, however, is that the welfare state does not stop people from using new tools to multiply their productivity. Anyone who punishes adaptation and rewards passivity will inevitably produce anger and a sense of entitlement.

Rising productivity, in return, opens up a space that had almost been lost: for charitable engagement and charitable consumption. Not compelled by the state, but voluntary. At best, people will move out of meaningless, wealth-destroying structures of redistribution into workshops of meaning, embedded in communal but voluntary orders.

Artificial intelligence does not guarantee this transition. But it makes it possible again.

Rahim Taghizadegan is the last Austrian representative of the Austrian School in the direct tradition, entrepreneur, author of more than fifteen books, university lecturer, and the founder of scholarium, citadel.garden, and deedsats.

Scholien by email · since 2009

Scholien. Monthly marginal notes by Rahim Taghizadegan on the present, written in the mode of the Viennese school.

One issue a month. Your address is never passed on; unsubscribe with every issue.