It’s perhaps clarifying to interpret generative AI as nothing more than an automation technology. Its purpose is to eradicate the perception that certain kinds of jobs require certain kinds of skills that workers possess, which give them potential leverage over their bosses, whose authority lies specifically in their lack of skill and their possession of capital.
“From the perspective of the capitalist entrepreneur the essential characteristic of the production process for which he is responsible is that it must operate itself,” Alfred Sohn-Rethel writes in Intellectual and Manual Labor (1977). “The controlling power of the capitalist hinges on this postulate of the self-acting or 'automatic' character of the labor process of production. This all-important postulate of automatism does not spring from any source in the technology of production but is inherent in the production relations of capitalism.” This is because capital’s authority derives from the fact of ownership rather than skill, or any specific exercised and embodied competencies. As long as capitalists control the production process, they can justify their extraction of profit at workers’ expense despite contributing no labor. “Automatic” is understood as the opposite of “artisanal,” as manufacturing dictated by the capabilities of machines.
Automation is not driven by new capabilities of new technology but by the demands of capitalism, which requires that workers be deskilled and disciplined. Under capitalism, machinery (and technology) must serve that purpose, and modern science is intrinsically oriented toward it. According to Sohn-Rethel, “The postulate of automatism as a condition for the capital control over production is even more vital than its economic profitability.” Just ask any money-losing gig-economy company. “The stages in the development of capitalism can be seen as so many steps in the pursuit of that postulate, and it is from this angle that we can understand the historical necessity of modern science as well as the peculiarity of its logical and methodological formation.” AI, then, appears as the current expression of capitalism’s “peculiar methodology,” the necessary form of “bourgeois science.”
Sohn-Rethel quotes Marx’s description of the factory from the “Machinery and Large-Scale Industry” chapter in Capital Vol. 1.:
Factory work … confiscates every atom of freedom, both in bodily and in intellectual activity. Even the lightening of the labour becomes an instrument of torture, since the machine does not free the worker from the work, but rather deprives the work itself of all content. Every kind of capitalist production, in so far as it is not only a labour process but also capital's process of valorization, has this in common … Owing to its conversion into an automaton, the instrument of labour confronts the worker during the labour process in the shape of capital, dead labour, which dominates and soaks up living labour-power. The separation of the intellectual faculties of the production process from manual labour, and the transformation of those faculties into powers exercised by capital over labour, is, as we have already shown, finally completed by large-scale industry erected on the foundation of machinery.
It’s not hard to interpret generative models and the data they depend on as allowing “dead labor” (the massive processing power aggregated by large tech companies) to “soak up living labor-power”(captured through surveillance as data) and turn it into a form of domination exercised over workers, conditioning their work processes as an increased form of dependency on capital. The prose that LLMs produce is always lifeless in this sense, by definition, regardless of its content. Reading that material is eating corpses, regardless of how it is seasoned. And it is easy to anticipate that “AI” will become an “instrument of torture” for workers, not only in threatening their livelihood but in “augmenting” their productivity by “confiscating every atom of freedom” in their activity and dictating the pace, decisions, and formulas to which they must adhere. Regardless of specifics, an LLM is issuing orders, training humans to subordinate themselves to its capabilities. Those capabilities are not an expression of what “new technology” can do; they are reassertions of capitalist relations of production and the power they have to impose themselves.
Another way to put that — and this idea that is implicit in every sci-fi depiction of rogue machines subordinating humanity — is that capital itself is always an “artificial intelligence” that emerges from the concentration of economic power. As Sohn-Rethel writes:
Capital is a social power which takes over production where it has outgrown the economic and technological capacities of the direct producer controlling it himself. While in the economic field the social power is capital, in the field of technology it is science, or, more accurately, the methodical operation of the human mind in its socialized form, guided by its specific logic, which is mathematics.
Mathematicization is what we might now call datafication, and contemporary surveillance can be understood in this light, as a means of gathering the data to allow for forms of labor to be simulated and brought under capitalist management protocols. In continuing his description of the factory, Marx writes that “the special skill of each individual insignificant factory operative vanishes as an infinitesimal quantity before the science, the gigantic physical forces, and the mass of labour that are embodied in the factory mechanism and, together with that mechanism, constitute the power of the ‘master.’"
This can be readily translated into contemporary terms. Generative models are factories; surveillance made all of us punch the clock. Our work vanishes into data, which the “gigantic physical forces” — the acres of servers harnessed by tech companies and their computer scientists — assemble automatically into the mechanisms that secure their power, dictating the only kind of labor that will be left for us to perform, that which reinscribes our own infinitesimal insignificance.
Internal exile is a reader-supported publication. To receive new posts and support my “work,” consider becoming a subscriber.
Delightfully fresh perspective and vivid metaphors. I will come back for more articles. But the abrupt end left me wondering what we should take away from this?
I don’t see any virtue in not exploring the capabilities of automation. After all I want to be free of traditional labour to engage in other endeavors.
We have reviewed your output and deem it provisionally acceptable.