When did you last check how your pension fund is invested?
When did you last ask which algorithm decided which news you saw first this morning?
When did you last think about why your supermarket arranges exactly these products in exactly this way — and what this arrangement does to your purchasing decisions?
Probably never. And that is entirely understandable. Life is short, attention is limited, and the automatisms work. The pension fund pays. The news informs. The supermarket supplies. So why check?
This is exactly where the subject of this essay begins.
The Friendliest Side of the Megamachine
The Megamachine — the anonymous system of structures, institutions, rules, incentives and automatisms that organises our society — is mostly described here as threatening. As a power that expropriates inventors, hollows out democracies, produces inequality. All of that is true. But it is not the whole story.
The most successful side of the Megamachine is its friendliest. It takes work off our hands. Effort. Decisions. Responsibility. It organises life so that one no longer has to worry about most of its basic conditions. The market sets the price — one does not have to negotiate. The algorithm chooses the news — one does not have to search. The financial system manages one's savings — one does not have to think. The insurance takes on the risk — one does not have to make provision.
That is not an attack. It is an extraordinary achievement of civilisation. Only a hundred years ago daily life required a great deal of work that is done automatically today. Water comes from the tap. Light comes from the socket. Food comes from the shelf. Money comes from the machine. That is liberation — liberation from physical labour, from anxiety about provision, from a large part of the basic existential tension that has shaped human life for millennia.
The Megamachine is the most successful convenience machine in history. And whoever finds a machine convenient does not ask how it works.
Who Asks How the Lift Works?
You press the button. The door opens. You go up. You do not ask about motors, cables, safety systems, maintenance intervals. That is reasonable. It would be a waste of time if every lift user had to understand the mechanics of their lift. The division of labour exists precisely for this — so that not everyone has to understand everything.
But the lift transports only the body. The automatisms of the Megamachine transport life — opinions, decisions, perceptions, values, political convictions, economic opportunities. And here the question of who does the maintenance, who determines the direction, who presses the button for others, is no longer a minor technical question. It is the central political question.
The algorithm that decides which news I see shapes my picture of the world. It is not neutral — it optimises for something. For attention. For time spent. For outrage, because outrage gets clicks. I do not notice this because the algorithm is invisible. I only notice that I am informed — and that feels good.
The market that sets the price of my labour determines what my time is worth. It is not neutral — it reflects power relations, historical imbalances, structural disadvantages. I do not notice this because the market is abstract. I only notice my wage — and whether it is enough or not.
The financial system that manages my savings decides in my name what happens to them — which companies they flow into, which projects they finance, which structures they strengthen. I do not notice this because the chain between my account and the money's final use is long and opaque. I only notice the annual return — and whether it is satisfactory.
The danger of the Megamachine lies not only in the fact that it compels us. It lies also in the fact that it relieves us — so completely and so pleasantly that we stop asking what it is actually relieving us of.
The Democracy Gap and the Helpful Automaton
In 2003 Ralf Dahrendorf coined a term that is more relevant today than it was then: the democracy gap. By it he meant spaces in which political decisions are made for which there are no democratic institutions. Spaces that have become diffuse — in which one is governed without being able to point a finger at governments.
What Dahrendorf does not name — perhaps because it was not yet so visible in 2003 — is why these gaps are so stable. Why they do not seriously trouble anyone. Why the democracy gap does not generate outrage commensurate with its size.
The answer is the helpful automaton. The spaces in which nobody decides democratically are the same spaces in which automatisms do the work. The financial market regulates. The algorithm sorts. The international institution coordinates. That does not feel like powerlessness — it feels like relief. The gap is invisible because it is convenient.
That is the most dangerous thing about the Megamachine: for the ordinary user it generally produces no recognisable victims at the moment it takes effect. The algorithm that shapes my opinion today produces no visible injustice — it only produces a slightly altered perception of the world. The financial system that today channels my savings into structures I did not want produces no tangible injury — it only produces a return from a source that would give me a bad conscience if I knew. The democracy gap that today makes a decision without my participation produces no sense of loss — it only produces the quiet fact of a world that has taken shape without my involvement.
What One Might Ask Oneself
This is not a call to paranoia. It is an invitation occasionally to interrupt one's convenience. Not constantly — that would be exhausting and would achieve nothing. But sometimes.
These questions have no simple answers. That is the point. The automaton is helpful precisely because it makes the question superfluous. Whoever asks the question stops treating the automatism as a matter of course — and begins to see it as what it is: a choice that an automaton has made for him.
The Paradox of the Helpful Machine
The Megamachine is paradoxical: the more helpful it is, the more powerful it becomes — and the less its power is noticed. A system that compels us generates resistance. A system that relieves us generates gratitude. And gratitude does not ask questions.
That explains why Dahrendorf's democracy gap does not close — not because nobody can see it, but because nobody has a sufficiently strong incentive to close it. The automatisms work. The lift moves. So why look into the engine room?
Why — that is a question everyone must answer for themselves. There is a reasonable answer that goes: no reason. Life is short, the engine room is complex, and most automatisms really are helpful. It would be exhausting and counterproductive if everyone constantly questioned everything.
But there is also another reasonable answer: because the direction in which the lift travels is not random. Because the buttons are labelled — but nobody labelled them deliberately. Because the automatisms take away the work — but also the decision. And because a society in which nobody looks into the engine room any more will at some point stop knowing where it is being taken.
The Next Stage — When the Automaton Decides for Itself
The helpful automaton has only limited intelligence. The algorithm waited until I scrolled. The market waited until I bought. The financial system waited until I paid in. I was still the trigger — even if I could not oversee the consequences. The question of who had programmed the machine was still a question I could in principle have asked.
The step from passive automatism to autonomous AI entity is not a gradual transition. The automaton reacted to my inputs. The autonomous entity knows my preferences, my habits, my mood of the moment — and acts on that basis before I have made any input. It no longer needs me as a trigger. It already knows what I want. Or what it wants me to want.
And so an uncomfortable question becomes an urgent one: who programmed the AI — and what goals does it pursue?
That is not a philosophical question. It is the same question as: who determines which country I live in, which rules apply, which opportunities I have. Only without an election. Without a parliament. Without the possibility of voting anyone out. The AI systems now emerging are not being shaped by elected representatives. They are being shaped by private companies — with their own interests, their own values, their own goals. Some of these goals are written down and publicly accessible. Many are not. None has been democratically legitimised.
That is the democracy gap of the next generation — larger than anything Dahrendorf saw in 2003. He described spaces in which nobody decides democratically. What is emerging are entities that nobody has democratically mandated — and that shape things nonetheless.
The disturbing question remains unanswered. Not because nobody has thought about it. But because the problem has categorical gaps for which the existing regulatory frameworks were not built. Democratic legitimation works through territory and citizenship. Autonomous AI systems operate globally, without territory, without voters, without the possibility of being voted out.
What remains is the question itself. And the refusal to forget it merely because it has no answer.
A question that nobody asks is never answered. A question that is asked at least creates the condition for an answer — at some point, under different conditions, with different instruments that perhaps do not yet exist. The helpful automaton would rather we did not ask the question. After all, it spares us the effort.
That is the only reason to write this essay: not to deliver a solution, but to keep visible a question that the Megamachine would rather keep invisible.