BadHabitus

← Back to Blog

Blog · Technology and political economy

AI and the Crisis of the Wage Form

Automation, socialized intelligence, and the politics of post-wage capitalism

Paul Adler’s argument about automation begins from a deceptively simple point: advanced production becomes more social as it becomes more automated. Automation does not merely replace direct human effort with machinery. It reorganizes work around systems, interdependence, monitoring, repair, coordination, and responsibility. The worker is no longer only the person who applies effort to material. The worker increasingly becomes the person who keeps a complex system intelligible, functional, and accountable.

In “Automation, Skill and the Future of Capitalism”, Adler argues that automation-induced trends in work undermine the commodity form of labor power and create the premise for post-capitalist forms of organization. The claim is not that automation automatically liberates workers. It is that capitalism develops technologies and forms of cooperation that strain against the wage relation through which it continues to organize labor. The more production depends on collective systems, shared knowledge, and distributed responsibility, the less plausible it becomes to treat labor as an individual commodity cleanly purchased on a market.

This is the point to hold onto when thinking about AI. Artificial intelligence is not simply another labor-saving device. It intensifies the socialization of production while presenting that socialized production in fetishized form. The model appears to write, reason, code, summarize, translate, design, and advise. But what appears as the intelligence of the machine is also the accumulated result of human language, public research, code, art, documentation, classification, moderation, user feedback, and the labor of countless workers distributed across the technical and cultural infrastructure of society. AI presents social labor as machine intelligence.

This is commodity fetishism in a new technological form. In Marx’s account, relations among people appear as relations among things. In AI, relations among writers, programmers, teachers, artists, scientists, moderators, users, engineers, and organizations appear as a relation between a user and an intelligent object. The interface gives the impression that the machine itself is the source of knowledge. The social relation disappears into the artifact. Intelligence, collectively produced and technically organized, returns as the property of the thing.

The point is not that AI is fake, or that the machine is merely a theatrical mask for human labor. The machine is real. The technical breakthrough is real. The computational infrastructure is real. The fetish consists in treating the artifact as the independent source of intelligence rather than as a crystallization of social relations, accumulated knowledge, technical systems, public investments, and organizational routines. The more convincing the artifact becomes, the easier it is to forget the society inside it.

This is why AI so directly undermines the wage form. The wage relation individualizes compensation even as AI reveals the collective character of production. The firm pays workers as isolated sellers of labor power, but the productive force it now seeks to exploit is not reducible to any individual worker. It is collective knowledge, socially accumulated and technically reorganized. The output appears to come from the model. The value depends on a social world. The private firm then claims ownership over the system that mediates this social intelligence and uses it to weaken the wage relation itself.

Adler’s argument helps explain why this is not merely a story of job displacement. Automation undermines the wage form because it makes work more complex, more interdependent, and less reducible to individually measurable output. AI radicalizes each of these tendencies. The individual worker’s contribution is harder to isolate because the work increasingly depends on systems. The value of the output is harder to attribute because the output condenses social knowledge. The justification for the individual wage becomes less stable because the productive force being mobilized increasingly exceeds the individual worker.

Capital’s fantasy is not simply to replace this or that worker. It is to break the connection between labor, income, and bargaining power while preserving private ownership over the machinery that now appears to produce intelligence. In this sense, AI does not threaten the wage relation from outside. It exposes the wage relation as inadequate to the level of socialized production capitalism has created. The worker is treated as individually replaceable at precisely the moment production has become more collectively dependent than ever.

The political contrast with China is revealing. Recent Chinese court decisions suggest an effort to prevent firms from using AI replacement as a unilateral pretext for destroying employment. A Hangzhou court ruled in favor of a worker dismissed after his employer claimed AI could perform his job, awarding compensation and treating the dismissal as unlawful. The case does not mean China has banned automation or prohibited all AI-related restructuring. It means Chinese legal institutions are signaling that firms cannot simply offload the costs of technological transition onto workers while treating AI as a managerial excuse for wage destruction.

That posture is strategically coherent. China wants AI development, but it also wants employment stability, effective demand, organizational capacity, and social order. The state’s position appears to be that AI should be used to increase productivity, not simply to give firms a legal fiction for dismissing workers. This is not socialism in the emancipatory sense, and it is not a solution to exploitation. It does show a state trying to discipline the terms on which automation enters the labor process. AI is not allowed to become simply the private right of the firm to dissolve the wage relation at will.

The American case is almost the inversion. In the United States, the most powerful capitalists increasingly speak as if the wage relation is becoming obsolete, but they do so while preserving private ownership over the productive apparatus. Elon Musk has described “Universal HIGH INCOME” through federal checks as the best response to unemployment caused by AI. The sincerity of this rhetoric is doubtful. Capitalists rarely discover the socialization of wealth at the exact moment they own the machines. But the concession is revealing. If AI proceeds according to the substitutionist logic favored by capital, then wages may no longer function as the main mechanism for distributing purchasing power.

This is post-wage politics without social ownership. The machine, the model, the data center, the cloud infrastructure, the intellectual property, and the asset value remain privately controlled. The public receives, at most, a compensatory payment after its bargaining power has been destroyed. This is not the socialization of wealth. It is the socialization of dependency after the privatization of productive capacity.

Bernie Sanders’s proposal for an AI sovereign wealth fund is important because it shifts the question from income support to ownership. In “The Public Should Own Half of the Big A.I. Companies,” Sanders argues that AI is built on collective human intelligence and that the wealth it generates should benefit the public. His proposed American AI Sovereign Wealth Fund Act would give the public a direct ownership stake in major AI firms through a one-time 50 percent stock tax, with voting shares and board representation. This proposal takes seriously the fact that AI encloses a social productive force. If the technology rests on collective intelligence, then the gains cannot legitimately be captured only by private owners.

But the proposal also reveals the limits of the American state. A serious AI sovereign wealth fund would require the state to discipline capital at the point of ownership. It would have to confront not only the cultural power of Silicon Valley, but the financial, infrastructural, military, and geopolitical interests now attached to AI development. The United States can subsidize chip production, procure AI systems, defend intellectual property, and protect asset values. It has far less capacity to subordinate the owners of AI to a democratic claim on the wealth AI produces.

Adler’s argument clarifies the stakes. AI is not merely a new threat to employment. It is an intensified socialization of production appearing in fetishized form as machine intelligence. The more social production becomes, the more absurd it is to organize income, authority, and ownership through individualized wage labor and private command. Yet that is exactly what capital is trying to do: use a socially produced technology to undermine the wage while preserving the property relation.

The task is therefore not to defend the wage form as if it were the horizon of justice. The wage has always been a compromised form, tying survival to labor-market dependence and subordinating collective productive capacity to private ownership. But neither can the destruction of the wage be left to capital. If AI weakens the wage relation, the political question becomes who controls the transition beyond it. Does post-wage society mean public ownership, shorter hours, stronger social provision, and democratic control over productive infrastructure? Or does it mean private AI monopolies, mass dependency, degraded work, and occasional checks from a state too weak to govern the firms it subsidizes?

AI sharpens the contradiction Adler identified. Capitalism socializes production, then tries to contain that socialization inside private property and wage labor. With AI, the contradiction becomes harder to hide. Intelligence appears as a commodity owned by firms, but its substance is social. The wage relation persists, but the productive force being mobilized exceeds the individual worker. Capital dreams of abolishing labor costs without abolishing private appropriation. The struggle now is whether the socialization of intelligence becomes a basis for democratic control or another enclosure of the common powers of society.

References