In this text, I will examine Jernej Markelj and Claudio Celis Bueno’s 2024 article, Machinic Agency and Datafication: Labour and Value After Anthropocentrism, through the lens of Marx’s concept of value. While the authors posit that data can generate value, I argue that non-human entities cannot participate in the production of value. According to Marx, value itself is a quantitative sign and functions as a semiotic system between humans. To substantiate this argument, I will first provide an in-depth analysis of Markelj and Bueno’s article and subsequently discuss the lack of relation between value and non-human entities based on Marxist theory.
The authors suggest that Marx’s political economy is characterized by “human exceptionalism” and believe this exceptionalism must end in the age of digitalization and automation (Markelj & Bueno, 2024, p. 1061). They strengthen this argument by claiming that capitalists exploit not only workers but also non-human actors, such as machines, during production. They assert that data can produce value, arguing that the path to this lies in “machinic agency.” Machinic agency implies that agency is not an intentionality unique to humans; rather, it emerges when human and non-human entities interact to enhance their capacity to act (Markelj & Bueno, 2024, p. 1066).
The authors attempt to invalidate Marx’s distinction between constant capital (machines) and variable capital (human labor) (Markelj & Bueno, 2024, p. 1063). They criticize the arguments of Caffentzis and Virno—who claim that machines cannot produce value and that capitalism instead exploits human cognitive and linguistic capacities—as being anthropocentric (Markelj & Bueno, 2024, pp. 1063-1066). For the authors, value is the production of capacities that human and non-human entities generate by coming together, which they could not achieve alone. In developing this explanation, they utilize Actor-Network Theory, which removes the human from the center of agency (Markelj & Bueno, 2024, p. 1066).
However, it must be noted that for the capitalist, non-human objects (i.e., technical machines) cannot be exploited because they are not entities that can be argued with or subjected to social conflict. They can, however, be appropriated. The authors overlook this distinction. The fact that data can be purchased with money does not mean it is a “value”; it means it is “valuable.” In a capitalist economy, these two are fundamentally different.
The core of my critique lies in the distinction that data is “valuable” rather than a producer of “value.” Labor power is brought together with technical machinery and raw materials in the space of production. A commodity cannot be value itself; rather, the “value” is the selling price that exists above the production cost of that commodity. To refute the “data is value” argument, let us construct a hypothetical scenario: Women’s menstrual cycles and symptoms are data points—meaning they are commodities. In the production of this commodity, physiological cycles, mobile phones, the fingers used to input data, and the time spent entering that data constitute the technical machinery and raw materials. The sum of these is the value—specifically, the money a third-party company pays to acquire this data. Here, non-human entities do not produce anything; they merely participate in production. This is not a rejection of anthropocentrism. As argued in the introduction and justified here, value is an independent sign regime—a semiotic system—imposed externally onto the interaction of animate and inanimate entities.
The authors claim Marx is anthropocentric, stating that he ignores nature and the wealth it creates. Yet, in Marx’s critique of the German Social Democratic Party’s Gotha Program, he states that nature is the direct producer and source of the resources that meet the needs of living beings: “Labor is not the source of all wealth. Nature is just as much the source of use values (and it is surely of such that material wealth consists!) as labor, which itself is only the manifestation of a force of nature, human labor power” (Marx and Engels, quoted in Karayemiş, 2026, p. 36 fn. 34).
The authors proceed with the premise that capitalists pay for algorithms or data. In reality, they do not pay for them; algorithms or data are appropriated. From a feminist perspective, is women’s domestic labor important to the capitalist? Certainly, but they respond to this importance by paying the employee a wage. This wage may not always cover both the male worker’s expenses and the woman’s domestic labor. In this case, the capitalist appropriates the woman’s reproduction. Similarly, in this context, data is not a value. It is merely an object of appropriation when it is purchased by third-party companies.
Markelj and Bueno’s attribution of data and machines as autonomous “value-producing” agents—drawing on Deleuze, Guattari, and Actor-Network Theory—indicates a misunderstanding of how capital functions. The argument I have established throughout this text is that data is not a “value” in itself, but a “valuable commodity” appropriated by capital during the production process. (I wish to emphasize its status as a commodity). If Marx were alive today, he would not view data as valueless. He would, however, criticize the commodification of the physiological cycles of the human body through datafication in digital capitalism. In short, data is not the thing that is quantitatively embodied and measured by money (value); it is the carrier of value.
REFERENCES
Karayemiş, O. (2026). Ekoloji ve Ekonomi I: Sermayenin Girişimleri ve Yeryüzünün Mukavemeti. başıbozuk e-yayıncılık.
Markelj, J., & Bueno, C. C. (2024). Machinic agency and datafication: Labour and value after anthropocentrism. Convergence: The International Journal of Research into New Media Technologies, 30(3), 1058-1075. https://doi.org/10.1177/13548565231166534