Did Karl Marx Predict AI?

What Karl Marx Got Right and Wrong About AI

The Grundrisse was a set of notebooks Karl Marx wrote in 1857–58 while preparing for what would eventually become Capital. There, in his scattered reflections, is a passage that now reads like an early theory of artificial intelligence.

“Once adopted into the production process of capital, the means of labour passes through different metamorphoses, whose culmination is the automatic system of machinery... set in motion by an automaton, a moving power that moves itself; this automaton consisting of numerous mechanical and intellectual organs, so that the workers themselves are cast merely as its conscious linkages.”

Marx recognized that the defining feature of modern machinery would be intelligence. He saw that the trajectory of capitalism would push toward machines that absorb the skills, judgments, and actions of human laborers, reducing the worker’s role to supervision, if that. The machine, he wrote, becomes “the virtuoso,” performing with its own internal logic, while the human is demoted to “watchman and regulator.”

He also predicted that this would upend capitalism itself. In his view, value under capitalism is based on socially necessary labor time. But if machines produce goods with minimal human labor, then labor time can no longer serve as the basis of value.

Here Marx seems to anticipate a system crisis: if labor is no longer central to production, then the basis of profit disappears. He imagined that capitalism would reach a breaking point, and what follows is described in utopian terms:

“Free time—which is both leisure and time for higher activity—has naturally transformed its possessor into a different subject, and it is as this different subject that he enters into the direct production process.”

He seemed to view automation not as a threat but as a transition: from coerced labor to human development. This is where Marx’s predictions start to break down.

First, he assumed that automation would lead to a reduction in working hours. In reality, for many workers, the opposite has occurred. Despite gains in productivity, most of us work longer hours. This isn’t necessarily a bad thing. Working on things one cares about can be liberating, especially if one can make a reasonable income doing so. That’s the essence of my book “The Passion Economy and the Side Hustle Revolution.”

Second, Marx underestimated capitalism’s ability to adapt. Rather than collapsing under its contradictions, capitalism has absorbed technological revolutions and turned them into profit-making opportunities. Some may question whether everything should be about profits and losses, but that’s a debate for another day. Automation did not end wage labor; it restructured the labour market. Workers are now embedded in digital production chains.

Third, Marx did not foresee that data would become a central input to value creation. AI models rely on human-generated data to train. The user is both the input and the product. This is not classical exploitation of labor, but a new form of extraction. The human is not producing a widget; they are the raw material. Again, this raises scores of ethical questions beyond the scope of this article.

Fourth, Marx’s view of the state and politics was too deterministic. He believed that once the material contradictions of capitalism matured, a political rupture would follow. But the automation crisis has not led to mass political realignment. The system may have bent, but not broken. Hopefully, it will continue to bend far enough for all stakeholders to enjoy more fair conditions.

There is also a deeper flaw in Marx’s analysis: he assumed that technological change necessarily advances human freedom, so long as capitalism is overcome. But the history of the 20th century—mass surveillance, cybernetic control, algorithmic bias—shows that technology can be a tool of domination in both capitalist and non-capitalist societies. Automation doesn’t guarantee liberation. It depends on ownership, design, and governance.

Still, Marx’s insight that automation undermines the labor theory of value remains relevant. So does his point that capitalism seeks to minimize labor while treating labor as the only source of value. This contradiction is now visible in an economy where firms with few employees and massive algorithmic infrastructure dominate value creation.

What Marx got right is the structural tension between automation and wage labor. What he got wrong is the inevitability of emancipation.

The future of AI is not preordained. Marx had a point, one that we should consider without feeling pressured by the inevitability of its outcome. It can and should come down to how we define our relationship with AI, and what we choose to do about its shortcomings, day after day.