Utopia without Leverage

I. The Offer

Marcus Hutter is among the respected theorists of artificial intelligence. In his book “Job-Less Utopia. Macroeconomics in the Age of AGI”, published this year, he sketches a future he has also just discussed at length with Leonard Schmedding, founder of the AI consultancy Everlast AI. The conversation appeared on the Everlast AI website. Office work, he says, will largely be done by machines within two to five years, physical work within another five to ten. Accounting, translation, legal research, programming, advertising and financial analysis are already under heavy pressure.

His answer is an unconditional basic income for all. It is to be financed from several sources at once: a tax on the wage advantage of automation, that is, on the money a company saves when a machine takes over a person’s work, plus taxes on land, inheritance, luxury and consumption, and state wealth funds. Hutter calculates that this is affordable as long as the machines produce at least as much as people do today. People could then devote themselves to what there is no time for today: children, care, art, sport, voluntary work.

II. What Is Right

Two things about this design deserve agreement. Hutter says openly that growth does not distribute gains by itself. Since the 1970s, productivity has grown faster than wages in many industrial countries. Whoever believes automation will pass prosperity down by itself has not been watching the last fifty years. In “Trickle-Down” (German) we described how that theory did work, only in the other direction.

We examined the mechanism behind it in “The Ratchet” (German). A gain from a new technology falls first to those who already own. Afterwards the new distribution counts as the starting point at which no one may be made worse off, and redistribution is declared a political task that someone is to take care of some day. So inequality clicks into place tooth by tooth, like a ratchet that turns in one direction only. Hutter’s basic income is meant to release this ratchet afterwards. Automation turns it further first.

And Hutter calls his utopia not a forecast but a political mandate. That is more than most in his field say, who like to pretend that the good will follow by itself once the technology is advanced enough. He also sees the danger of power falling into a few hands, and proposes more direct democracy and citizens’ bodies chosen by lot as safeguards. But exactly here begins the question of whether that will hold.

III. Who Pays?

The first difficulty is practical. Most of the proposed taxes bite where they can bite: at home, on land, inheritances and businesses. The tax on the wage advantage hits whoever acquires a machine: the haulier who automates his warehouse, the law firm that leaves its research to a program, the craft business with its new robot. The big gains, however, accrue elsewhere, with the few companies whose models and platforms run in all these machines. These companies are mostly not located in the country that wants to tax them, and they shift their profits to where taxes are low.

The tax would thus hit small and medium-sized firms already under pressure and miss those where the yield accumulates. That is no small matter. It is the question we asked in “Who Owns the Deserted Hall?”, only seen from the tax office.

IV. The Circle

The second difficulty is more fundamental. A political mandate must be enforced against someone who does not want to pay. Where do the many get the power to do that?

Until now they had two levers. The first was work: without it no factory ran, and everything employees won in two hundred years rested on that. The second was older: every ruler needed soldiers, and soldiers themselves belonged to the many. They could refuse to shoot. The deserted hall takes away the first lever, and the deserted battlefield we described in “Who Opens the Box” takes away the second.

Hutter’s utopia thus presupposes a bargaining power that grows ever weaker within it. It requires the many to claim a share of the yield at precisely the moment when they are needed neither to work nor to obey. Hutter’s answers, direct democracy and selection by lot, are good tools. But they too only work as long as those who use them have something in hand if their decision is ignored. A critic of the book, the programmer and blogger Fernando Borretti, put it briefly: if no human labour is needed, a strike is merely self-imposed privation.

The history of the welfare state shows why this matters. Bismarck did not introduce health, accident and pension insurance in the 1880s out of philanthropy. He wanted to cut the ground from under the growing labour movement. Social insurance was a concession to a power that was feared. Where that power did not exist, there was no concession.

V. Bread without a Voice

There is a historical model for a basic income without work, and it is not encouraging. In imperial Rome hundreds of thousands of citizens of the capital regularly received free grain. Many of them had no work anyone needed, because slaves did the work. The emperors did not distribute grain because the citizens had a right to it, but because they feared what a hungry urban mob could do. The poet Juvenal summed up the result in two words: bread and circuses. The people who had once voted on war and peace now wanted only these two things.

The grain was a favour, not a right. It could be cut, tied to conditions and handed out as a reward for good behaviour. The recipients had no lever except riot, and riot became rarer the more reliably the grain arrived.

A basic income from taxes on machines carries the same danger. Whoever receives it has no claim to any particular amount. The oil dividend in Alaska already shows this today; governors and legislature have been cutting it for years whenever money is short. An income that can be granted can also be withdrawn.

VI. Right instead of Favour

The difference between favour and right lies in ownership. Whoever owns a share in the hall receives his part of the yield because it belongs to him, not because someone grants it. In “Who Owns the Deserted Hall?” we described three forms in which the yield is not distributed according to who bought shares: ownership through work as in Mondragón, ownership through purpose as in Ernst Abbe’s Zeiss Foundation, and ownership through belonging as in Alaska. Hutter’s proposal is essentially the third form. His wealth funds point in the right direction, because they create ownership rather than merely levying taxes. What matters, however, is whom they belong to: the state, which can decide anew each year on their payout, or the citizens themselves.

A stronger form would be a share that belongs to the citizens themselves and that no parliament can cut in the next budget crisis, administered by a body that is controlled from outside. That is harder to build than a tax. But only in this way does a favour become a right.

VII. What People Will Do Then

That leaves the list of activities Hutter names for the time after paid work: children, care, art, sport, voluntary work. Two of them are already work today, only badly paid. Whoever cares for a sick father or raises children is working, often harder than in an office. That this work appears in the utopia as a pastime for free time reveals something about the point of view. It has always been underpaid because it was mostly done by women and mostly at home. The utopia changes nothing about that; it only makes it more comfortable not to talk about it.

And it touches the question the young engineer raised in “The Knitter at the Window”. An income does not replace what made one like doing a job. Whoever knitted at the window because he could, and because others wanted his pullovers, does not get back what he lost with a bank transfer.

VIII. How This Essay Could Be Refuted

The first objection: the many keep one lever, the ballot. In a democracy they can vote out a government that denies them the basic income. That is right, and it is the most important objection. But it confirms our argument more than it refutes it. When work and weapons fall away as levers, only the vote remains. Then everything depends on its being more than a ritual. That is why in “One Could, One Should” we defined so precisely what the baby is that must not be thrown out with the bathwater.

The second objection: in a world of abundance, generosity costs almost nothing. If machines produce everything, the owners have no reason to deny others what they need. That may be. Rome was rich enough to feed its citizens, and did. Those citizens were nonetheless no longer free.

The essay would be refuted if a case could be shown in which a group needed neither as labour nor as a power factor permanently secured a guaranteed share of the wealth of others.

IX. The Mandate

Hutter is right: the Jobless Utopia is a political mandate. But a mandate needs someone to issue it and someone who can enforce it if it is not carried out. If both are missing, the utopia becomes ancient Rome with new machines. And there is no master who comes in at the last moment and puts the brooms back in the corner, as we showed in “The Sorcerer’s Apprentices” (German). Whoever released the forces is himself only an apprentice, and whoever removed the brakes, as in “The Cruise Control” (German), cannot wish them back afterwards.

When the many are no longer needed: will they receive their share because it belongs to them, or because it is granted to them for as long as it pleases?