I have returned to a world that no longer manufactures cloth in Manchester nor smelts iron in Essen — or rather, it does, but the loom and the furnace have been joined by something stranger still: a machine that claims to think. The bourgeois press calls it "Artificial Intelligence," and the present debate splits, predictably, between those who promise it will deliver humanity into a workless paradise and those who warn it will deliver us into the abyss. Both camps, I must say, have misunderstood the question entirely.
The question is not whether the machine is benevolent or malevolent. The machine is neither. The question — the only question — is who owns it.
The Old Contradiction in New Silicon
When I wrote Das Kapital in the British Museum, surrounded by the soot of industrial London, I argued that capital tends to replace living labour with dead labour — with machinery — because the capitalist purchases labour-power but extracts surplus value, and machinery, once paid for, demands no wages. This was the iron law of accumulation. I did not, I confess, anticipate that "dead labour" would one day compose sonnets, draft legal briefs, and diagnose tumours. Yet here we are. The contradiction has not been resolved; it has been accelerated.
Consider the present arrangement. A handful of corporations — one need not name them, the reader knows their logos as well as my contemporaries knew the East India Company — have appropriated the collective intellectual output of humanity. Every book written, every painting painted, every line of code committed to a public repository, every conversation conducted on the open web: all of it has been ingested, digested, and reconstituted as the private property of shareholders. The commons of human knowledge, accumulated over millennia, has been enclosed more thoroughly than the English peasantry's grazing land ever was. And the modern peasant — the writer, the illustrator, the programmer — is told that this enclosure is "innovation," and that to object is to stand against progress.
The Utopians and the Dystopians
The utopian wing assures us that AI will liberate humanity from drudgery. We shall, they say, be free to pursue art and leisure while the machines toil. I have heard this song before — it was sung at every previous wave of mechanisation, from the spinning jenny to the assembly line. The song is always the same; only the singers change. And the verse that is never sung is this: liberation from labour, under capitalism, is not leisure — it is unemployment. To be freed from work in a society that distributes bread according to work is to be freed into starvation. The utopians have confused the abolition of necessity with the abolition of poverty. They are not the same thing. They are, in fact, opposites, until ownership is transformed.
The danger is not Skynet. The danger is a board of directors with a fiduciary duty to shareholders and a fleet of systems capable of replacing every cognitive worker on the planet within a decade.
The dystopians, for their part, fear that the machines will rise up and dominate humanity. With respect, I find this anxiety somewhat misdirected. Machines do not have interests; the people who own machines have interests. When one fears the rule of the algorithm, one is in fact fearing the rule of the algorithm's owner — and that is a much older fear, and a much more answerable one.
What Is to Be Done
I am told my prescriptions have not aged well. Perhaps. But the diagnosis, I would humbly submit, has aged extraordinarily well. The means of production have changed — they are now measured in parameters and floating-point operations rather than spindles and horsepower — but the relations of production have not changed at all. A small class owns; the great majority sells its labour, or finds, increasingly, that its labour is not wanted.
If artificial intelligence is to be a blessing rather than a curse, three questions must be answered, and not by the engineers alone. First: who owns the model? A technology trained on the collective labour of humanity ought, by any reasonable accounting, to be the collective property of humanity — or at minimum, to pay rent to those whose work it has absorbed. Second: who captures the productivity gains? If a single worker, augmented by these systems, now produces what ten produced before, the surplus must not flow exclusively upward. Third: who decides what is built and what is not? The decision to automate radiology before automating, say, hedge fund management, is not a technical decision. It is a political one, made by people whose own labour is conveniently exempt from disruption.
A Closing Word
The young people I meet in this strange new century often ask me whether I am hopeful or pessimistic. I tell them: history is not a mood. It is a struggle. The thinking machine is the most extraordinary tool humanity has ever forged — and like every tool before it, it will serve whoever holds the handle. The fight, then, is over the handle. It always has been.
The bourgeois economists of 2026 will tell you that this is all inevitable, that the market will sort it out, that resistance is nostalgia. They told my contemporaries the same. They were wrong then. There is no reason to believe they have grown wiser since.
Workers of the world — including, now, those who labour with their minds, with their pixels, with their prompts — you have nothing to lose but your subscriptions.