Notes on the Political Economy of Artificial Intelligence

Ahmet Cengiz

This article was originally published in Teori ve Eylem, No.72, June 2026

Starting some time before, but especially since 2022, a veritable artificial intelligence (AI) craze began—or rather, was set in motion—following the emergence of generative AI (Gen. AI) systems like ChatGPT, i.e., algorithms capable of independently generating text, images, or code. On the one hand, this craze has led to a massive increase in investment in AI technology—and even, in terms of volume, to the formation of a bubble in the stock markets. For example, as these lines were being written, the following news broke: “Surge in demand for AI chips has quadrupled Samsung Electronics’ stock price over the past year, pushing the company’s market value above $1 trillion.” Thus, Samsung has become the darling of chip manufacturing and the second Asian tech giant—after Taiwan’s TSMC—to reach this level, with the global chip market projected to grow to $1.5 trillion by 2030. [^[1]] Of course, the real giants are the “Magnificent 7” in the U.S.—namely, Apple, Microsoft, Alphabet (Google), Amazon, Nvidia, Meta, and Tesla. They truly are “magnificent,” as together they account for approximately 35 percent of the U.S. stock market! With a total investment volume of approximately $560 billion, they are also the primary driving forces behind the investment wave in generative artificial intelligence.

Also, the frenzy has brought about a resurgence of various fantastical views—similar to those we’ve seen in the past regarding technology fetishism (such as the automation debates of the 1950s and 1960s)—but this time, they seem to be gaining traction as if they were more realistic. Both utopian (a “golden age” of abundance where work will disappear thanks to smart machines and no one will need to work) and dystopian (a society where, as a result of AI, both human abilities and humanity itself will become obsolete) views have resurfaced. However, despite the increasing diversity of these views—and even though they may appear to stand at opposite ends of the spectrum—their common ground has remained largely unchanged. This common ground, which essentially stems from technological determinism, is the assumption that technology is a neutral yet transformative agent—or, in effect, a devil possessing god-like power. As is well known, in this approach, a specific history and society—along with their specific material conditions and class-based social relations and struggles—are external to technology; or, to put it another way, technology is an entity in and of itself, abstracted from these factors.

Undoubtedly, every technology has its own unique structure, a potential—whether mobilized or not—inherent in that structure, and, in terms of the totality of these characteristics, corresponds to an evolution or progress, and thus to a history. This reality, however, does not point to the autonomy of that technology, but rather to the material and social relations within which it is conditioned and in which it flourishes and develops. Indeed, every technology that has played an active role in social relations and life has not only corresponded to specific material and social needs or purposes but has also been the subject of class struggles within that society. The same applies to AI technology itself and the craze surrounding it.

If technology cannot be evaluated independently of the relevant economic and social formation, then this reality implies that AI could be used in a socialist society for different purposes and in different forms than it is today. It is clear that this technology offers new possibilities for a socialist economy and society.

What is Artificial Intelligence, and what it is not?

However, before addressing this topic, it is necessary to briefly outline some key aspects regarding the uniqueness and architecture of AI technology. The concept of AI itself, along with terms used in that context—such as “artificial neural networks,” “machine learning,” or “deep learning”—are actually misleading. Such anthropomorphic definitions suggest that AI, by mimicking the functioning and neural network structure of the human brain, will (or has) perform cognitive abilities, primarily intelligence. In reality, however, there is neither intelligence nor thinking nor neural networks in the true sense of the word. What we are dealing with is, rather, a system that reshapes the static structures in the data used to produce the desired new outputs. Here, it is less a matter of “understanding” and more a matter of transforming “learned” probability patterns into artificial responses. Human intelligence, however, cannot be reduced to merely the ability to process data or correlate patterns. Intelligence stems from the social practices of people embedded in concrete material relationships, from labour that transforms nature and is in turn transformed by it, from language that carries and conveys shared meanings, and so on. Setting aside the fact that humans possess self-awareness and rationality, are capable of making thinking the subject of their own thought, and have a historically shaped understanding of the world and their place within it—none of these exist in AI.

To put it in terms of two examples: AI is “the scientific discipline that investigates how we can get artificial systems (which may have physical bodies) to perform any cognitive activity (whether intelligent or not) that natural systems are capable of, at even higher levels of performance.”[^[2]] Or: “AI, quite simply, is the practice of getting computers to do the kinds of things that minds do.”[^[3]] This “getting them to do,” however, essentially consists of an abstract compression of past human labour, experience, and language into data. And this occurs as a digital reproduction of what essentially already exists, without any connection to or practical engagement with the real contradictions of material life and nature.[^[4]] As noted in a critique of large language models (LLMs) like ChatGPT, AI systems are actually “stochastic parrots.” [^[5]] Just as a parrot does not understand the responses it gives, these systems reproduce the views of the ruling class and the racism, sexism, and other biases embedded in the data used to train them. At this stage of their development, they convey knowledge of only a limited period in human history.

Although we speak of AI technology or technologies within the framework of the standardization of knowledge—where information is increasingly converted into structured, machine-processable data units—it must be emphasized that the technology before us is, in its latest form, a technology integrated into information and communication technology based on digitalization, encompassing electronics, microelectronics, semiconductors, chips, integrated circuits, computers, the internet, massive amounts of big data, etc. in short, a technology integrated into information and communication technology based on digitalization in its latest form. It can also be expressed as follows: AI is a specific dimension of digital technology that operates on the basis of algorithms and data. In other words, AI is not a technology in and of itself but gains its technological distinctiveness within the specified technical foundation and context. For example, from the perspective of AI algorithms, high-performance chips, processors, and storage units are required for digital big data to be functional. Without these, the extraordinary computational speed that makes the fundamental essence of AI technology possible would not be achievable. Such highly advanced hardware components, in turn, require the production of various rare and valuable raw materials through sophisticated manufacturing processes. This speed also requires massive amounts of energy (electricity) and water. Clearly, not every country or company can possess a technology whose technical infrastructure, architecture, and operations rely on such a broad spectrum of resources. In this regard, while AI may appear to be accessible to everyone, it is, in terms of its context and infrastructure, an extremely monopolistic technology. And we believe there is no need to point out that monopolization in this field is extraordinary, with a few technology monopolies—led by the U.S.—calling the shots.

Undoubtedly, AI represents a new phase in information and communication technology with its capabilities for rapid calculation of possibilities and cognitive inferences based on patterns. While this is one aspect of the matter, the other is that the requirements for the formation of AI’s own technological uniqueness encompass a broad field and, consequently, carry the potential to cause disruption across a wide range of areas (some of which are already evident). The contextual nature of material relationships—particularly the emergence of the ability to transfer them to the digital realm as a result of advancements in microelectronics and digital technology—and, furthermore, the fact that industry today cannot maintain its functionality without digital technologies, means that AI presents both a significant opportunity and a major threat. By facilitating access to information, undertaking certain cognitive processes, performing some time-consuming yet indispensable tasks, creating models and blueprints across a wide variety of subjects and fields, and identifying contextual relationships and patterns, among other capabilities, AI serves as a major facilitator and, in particular, a significant time-saver.

However, in order to deliver the expected efficiency in every field, it must not only objectify and standardize information—that is, it must not destroy the subjectivity of knowledge and thereby further deepen the alienation of labour in this regard—but it also, by its very nature, tends toward totality in knowledge. This tendency manifests not only in the comprehensive coverage of all data transferred to the digital environment but also in the production process itself. The use of labour’s mental capacity (its knowledge and experience) in the capitalist mode of production is, of course, not new (e.g., Taylorism).[^[6]] Digitalization, the internet, and AI technologies are bringing about a shift in which the knowledge of an individual or a group of workers is generally being replaced by accumulated process knowledge. Here, the measurement, evaluation, and—consequently—direction of knowledge regarding the labour process are increasingly being undertaken by automation mechanisms linked to digital networks. In the capitalist mode of production, which is based on the exploitation of living labour, while full automation of overall production is not possible, it is possible to centralize data from certain stages of the labour process in the manner described. And wherever this is possible (as is the case at Amazon or Walmart, for example), workers are reduced to serving as the sensory-motor extensions of digitally directed and processed systems.

Today, the vast amounts of data [^[7]] on economic, social, and cultural life that have been transferred to the digital realm constitute the sphere of activity for algorithms produced and shaped in the interests of one technology monopoly or another, without any social or public control mechanisms worth mentioning! Of course, many speculative claims can be made about “technological singularity”—a concept associated with transhumanism and promoted by certain spokespeople for U.S. tech monopolies—namely, “super AI” (ASI)—as promoted by certain spokespeople for U.S. tech monopolies—can be debated, but setting aside the physical and economic limits to its feasibility, what truly demands attention are the multifaceted damages and dangers posed by today’s extraordinary monopolization of digital infrastructure, data, and information—extending from people’s private lives to their social status. Among these dangers, the placement of AI at the service of the capitalist state’s security and intelligence apparatus—particularly the military—and the development of cyber and autonomous weapons systems through this means hold a special place.[^[8]]

Another dimension of digital and AI technologies that is significant from the perspective of the capitalist class is their role in reinforcing ideological and cultural hegemony. In this regard, AI is not merely a guiding technology in a technical sense, but also a fully controlling technology from an ideological perspective. It offers the ruling classes unprecedented opportunities to distort the truth, spread lies and demagoguery, create illusions, and manipulate public opinion, while also playing a unique role in binding people to the virtual world and deepening alienation. Since the effects of these technologies—which can be turned into tools for such ideological functions—on people’s social relationships, cultural life, and creativity, particularly communication, are beyond the scope of our discussion, we will not address them here. Similarly, another issue we will not dwell on—though no less important—is the destructive impact these technologies, as tools of daily life, have on human cognitive abilities, particularly the capacity for abstraction, especially among young people. [^[9]] In a capitalist society based on competition and organizing education according to the needs of capital, it cannot be expected that measures to seriously mitigate the “side effects” of AI in these areas will be taken spontaneously. Both the implementation of such measures and the limitation of the impact of the ideological functions that AI performs for the ruling class can only be achieved through struggle.

Artificial Intelligence as a Digital Machine

In technology fetishism—which treats technology by abstracting it from the given mode of production in which it emerges and is applied, or more precisely, from the economic-social formation—there is undoubtedly a role played by the fact that technology, in a general sense, constitutes a constant in social production, such as in production and consumption. Indeed, technology is not exclusive to capitalism. But just as the type of social formation varies depending on the mode of production, so too do technology itself, its development, and its form change accordingly.

The prevalence of technological fetishism in capitalism stems from the fact that the more concrete reality that fuels it is specific to the capitalist mode of production. Marx expresses this reality as follows: “Modern industry never regards the existing form of a production process as a definitive form, nor does it treat it as such. This is why, although it is conservative in terms of the essence of all previous modes of production, its technical foundation is revolutionary. Through machines, chemical processes, and other methods, it constantly transforms not only the technical basis of production but also the functions of workers and the social composition of the labour process.” [^[10]]

What makes the technical foundation of modern industry revolutionary is the distinct role that the capitalist mode of production plays in the production of surplus value as the source of capital accumulation. As Marx demonstrated, while absolute surplus value is always present in the capitalist mode of production, its true originality lies in the production of relative surplus value. The fundamental lever for this is the increase in the productivity of labour. By the rise in labour productivity, Marx means “a change occurring in the labour process that shortens the socially necessary labour time for the production of a commodity—that is, the ability of a given quantity of labour to produce a greater amount of use-value.” [^[11]]

In capitalism, the development of the productive forces is not an end in itself. In Marx’s words, “the aim of the development of labour productivity in the capitalist mode of production is to shorten that part of the working day during which the worker is compelled to work for himself, so that the remaining part of the working day—during which the worker works for the capitalist without compensation—may be extended.”[^[12]]

Undoubtedly, there are many factors that determine the increase in the productivity of labour [^[13]], but it is clear that the most important of these—depending on “the level of scientific development and the technological availability”—are machines. From the perspective of capitalist production, technological innovations and the application of new technical possibilities to production depend directly on their contribution to or impact on the productivity of labour. In capitalism, “for this reason, the degree of productivity provided by the machine is measured by the amount of human labour power it replaces.”[^[14]] Therefore, the essential criterion for capital in the use of machinery is “the difference between the value of the machine and the value of the labour power it replaces.”[^[15]]

Capital decides whether or not to use more advanced machines or new technological innovations in the production process based on this calculation of difference. Therefore, what matters is not the mere existence of this or that technological innovation or technical capability. What matters is the cost of using them—or, more precisely, whether their use will increase the production of surplus value. This use of machinery also demonstrates how, under capitalism, technological possibilities are limited by a socially narrow-minded objective; consequently, the conditions for the development of science and technology—as significant productive forces—and their use for the benefit of society are undermined, thereby widening the gap between the potential and actual realization of this productive force.

Marx provides very striking examples of this in *Capital*. First, he points out that “the division of the working day into necessary labour and surplus labour varies from country to country.” He also notes that “it differs within the same country at different periods or, within the same period, across different branches of industry.” Furthermore, he states that “the worker’s actual wage may sometimes fall below the value of the worker’s labour power and sometimes rise above it.” Due to such factors, “the difference between the price of the machine and the price of the labour power it replaces may vary.” And in this regard, he draws attention to the following interesting situation: “For this reason, in England today, machines are being invented that are used only in North America; just as in the 16th and 17th centuries, machines were invented in Germany that were used only in the Netherlands, and just as certain French inventions of the 18th century were utilized only in England. In more developed countries, when machines are used in certain industries, they create such a surplus of labour (Ricardo calls this the “redundancy of labour”) that the fall of wages below the value of labour power prevents the use of machinery in these sectors; for the capitalist, this makes such use unnecessary—and in many cases impossible—since profit arises not from a reduction in the labour employed but from a reduction in the labour for which payment is made.”[^[16]]

Marx also draws attention to the opposite situation: “As soon as the working class’s growing rebellion forces the state to forcibly shorten the workday and, first and foremost, to establish the normal workday in factories in the true sense as the legal workday—and as a result, the avenues for increasing surplus value production by extending the workday are definitively blocked— capital, from that moment on, with all its might and consciousness, set about producing relative surplus value by accelerating the development of the machine system at an ever-increasing pace.”[^[17]]

The context Marx highlights here has not changed. For example, it is well known that over the past 30 years—a period characterized as “neoliberalism,” the essence of which was capital’s (successful) general assault on the working class—there was no significant acceleration in investment in advanced capitalist countries; nor was there a major leap in productivity, nor were there large-scale investments in expensive new industrial technologies. Yet it was precisely during this period that there were major increases in the accumulation of capital. And rather than increasing its investment in industrial technologies, capital turned much more heavily toward government bonds and stocks than before, or increased its efforts to take over other companies, while also generously distributing dividends to its own shareholders, and so on. However, there was another characteristic of this period: workers in advanced capitalist countries experienced serious declines in their real wages; in other words, the value of labour power was increasingly reduced through various means (by reorganizing the labour process, imposing flexible working conditions, de-unionizing to weaken labour’s bargaining power, etc.). At the same time, privatizations—particularly in sectors ranging from healthcare to education—opened up new areas of value creation for capital, leading to the expansion of the service sector, where wages are relatively low. This correlation is surely no coincidence!

According to Marx, fixed capital, in its physical form, undergoes constant transformation—from tools, through simple machines, to a “system of machines.” In this sense, AI—as a digital machine—represents a contemporary form of this transformation. Given that it is a fact that, within the capitalist mode of production, machines serve to increase surplus value by intensifying labour and reducing the necessary labour time, and considering that AI is also a digital machine operating through algorithms, it follows that today AI primarily serves to: on the one hand, generally devaluing the labour force by making it less skilled, and also intensifying labour and increasing the workload by enhancing capital’s control over labour (see the article titled “Artificial Intelligence, Digital Technologies, and Labour Control” in the dossier), on the other hand, in accelerating the rate at which surplus value is realized (for example, by shortening the circulation time of capital through digital platforms).

Consequently, in the sense described, digitalization and AI represent the ongoing transformation of fixed capital, this time characterized by the collection, storage, and evaluation of information. In short, what is changing are, in fact, the methods by which capital exploits labour, the techniques used to organize this exploitation, and the means by which it realizes its profits. Undoubtedly, from the perspective of capital, these new possibilities make its task somewhat easier, both during times of crisis and in overcoming the restrictive effects of the law of the tendency of the rate of profit to fall.

The points highlighted in the above quotations from Marx also shed light on the fact that AI is currently being used to varying degrees across different sectors, and that its impact—for example, as a tool for increasing labour productivity—differs significantly from sector to sector. For instance, while AI is far more effective in terms of labour productivity on digital platforms, the same cannot be said to the same extent regarding industrial production. At this stage, AI’s contributions to existing productivity in industrial production remain limited. Clearly, this has more to do with the impact of the elevated organic composition of industrial capital on profit rates than with the nature of AI technology itself.

In this context, the point that must be emphasized is that, as Marx emphasized, machines do not produce surplus value; they merely transfer the value they contain to the product in stages. This point is important because it makes it clear that the fantastical views put forward in the context of both robots and AI are not grounded in reality; the aim of capitalist production is to increase capital accumulation, and therefore surplus value cannot be produced without exploiting living labour—it can only be produced through dead labour (machines, and thus robots as well). This is an inevitable consequence of capital’s inherent contradiction: it is compelled to enhance the productive power of labour through machines, yet can only extract surplus value by exploiting living labour.

Artificial Intelligence and a Debate That is Far from Over

Undoubtedly, the final word on AI technologies has not yet been spoken. New technologies—such as “machine learning,” “large language models,” and “generative AI”—may be added to AI’s existing repertoire, and the weaknesses of current technologies may be addressed. It is well known that there is fierce competition between monopolies and states in these areas. Therefore, the points we have emphasized regarding machines and AI do not imply that there will be no technological progress in this field, but rather that technology itself—that is, its progress and application—is constrained by the capitalist mode of production; that its advancement and implementation are conditioned by a narrow objective (maximizing profit); and that this, in turn, leads to the negative consequences described above, particularly in the working lives of the labouring masses.

The fact that social production is conditioned by a narrow objective brings about significant differences between the capitalist and socialist modes of production, both in terms of technology in general and the use of machines in particular. In the quote we cited above regarding the use of machines, we saw that Marx stated the primary criterion for capital in the use of machines is “the difference between the value of the machine and the value of the labour power it replaces.” In both the first part of the sentence where he expresses this and the one preceding it, he touches on another important point: “If viewed solely as a means of reducing the cost of the product, the use of machinery is limited as follows: the labour expended in the manufacture of the machine must be less than the labour saved by its use. However, for capital, the limits of use are even narrower. Since what capital pays for is not the labour expended but the value of the labour power used, the use of machinery, for capital, is limited by the difference between the value of the machine and the value of the labour power it replaces.”[^[18]]

In other words, “even if the difference between the amount of labour required to produce the machine and the total amount of labour it replaces remains the same, the difference between the price of the machine and the price of the labour power it replaces may vary.” [^[19]] (The capitalist, compelled to increase profits and cope with the harsh laws of competition, focuses on this price difference.) It is precisely at this point, where Marx draws attention to this difference, that he adds the following footnote: “For this reason, the scope of machine use in a communist society would be entirely different from that in a bourgeois society.”[^[20]]

As we can see, the interrelationship between technology (and its use) and the economic and social formation reemerges: In a communist (and, as its first stage, socialist) society, the development and use of technology are not subject to the constraints of the capitalist mode of production, because the purpose of production is not to enrich a handful of capitalists who own the means of production, but to meet the needs of society. If we disregard vulgar bourgeois demagogues, neither friend nor foe can deny this fundamental difference.

The debate, however, stems from another point. This is the claim that a planned socialist economy cannot succeed, since a socialist economy cannot fulfil the functions that the “free market” and prices perform in the operation of the economy under capitalism; rather, by attempting to perform these functions through central planning, it will create a “command economy” that disrupts the economy’s internal balances. The source of inspiration for these claims lies in the debates that began in the 1920s under the names “The Economic Calculation Debate” or “The Socialist Calculation Debate”—debates in which the most fundamental arguments that bourgeois economists still use today against a socialist economy were first put forward.

We will not address this significant debate, which spanned decades, here; that would require a separate article. Nevertheless, it is worth briefly summarizing this debate, as the contributions that digitalization and AI technologies could make toward solving certain problems of a socialist economy also possess characteristics that refute the debate’s main argument.

In this debate—initiated by Ludwig von Mises, one of the founders of the “Austrian School,” and continued by his student Friedrich August von Hayek, who later gained fame as a theorist of “neoliberalism”—one of the central theses they defended was Mises’s argument that economically rational decisions are predicated on prices. Prices, in turn, depend on the effectiveness of the market mechanism, and the market, in turn, requires private ownership of the means of production. However, where there is no price mechanism, the prices of goods and services cannot be known, and therefore, rational calculation is impossible. Hayek, recognizing the weakness in the argument (“Prices make rational decisions possible, but why must they be necessary?”), sought to address it as follows: A capable actor can make rational decisions even without prices, but a real actor operating within the market—subject to the “flaws of human judgment”—will need prices to determine the right course of action. [^[21]] Since the information derived from prices cannot be centralized in a socialist economy, the Central Planning Board will face serious problems regarding pricing. For information is dispersed throughout the market; even if the products for which prices are to be set were identified and the necessary information to establish the supply-demand equation were gathered from the market, this would not solve the problem—since the supply-demand situation in the market is constantly changing, the information regarding these factors will also change; in other words, the fluidity of the market will render the information gathered at a specific point in time obsolete.

The arguments they put forward in this debate underwent some changes—not only due to the effective objections raised by Marxist economists but also because the socialist construction in the Soviet Union gained momentum during those years and the 1929 Great Depression shook the capitalist world—but their core thrust remained unchanged: namely, the thesis that because there is no free market in socialism—or, more precisely, because there is no market for “means of production” and factors of production—the value of these things cannot be objectively determined; and that, as a result of the absence of such a determination through the market process, costs (as a given) lose their meaning, and therefore meaningful economic accounting is impossible.

Marxist economists, led by the British economist Maurice Dobb, essentially defended the following idea in this debate, which went through various phases: It is possible to conduct rational calculations in a socialist economy without abandoning central planning and without allowing for the potential distortions that the compromise formula—proposed by theorists of market socialism such as Oskar Lange and others, based on the dichotomy created between the market and planning—might lead to. Dobb argued that the Mises School, in this debate, overshadowed the far more significant aspects of a socialist economy (changing relations of production, production, productivity, etc.), and claimed—based entirely on an a priori argument—that there is a fundamental incompatibility between social ownership of the means of production and rational economic calculation; but he stated in many of his articles that this argument had been clearly refuted throughout the debate. He noted that what the Mises School viewed as an a priori incompatibility was in fact merely a technical problem, that Marx and Engels had drawn attention to this problem, and that its solution depended on refining the organization of central planning and implementing technical measures.[^[22]] In other words, the debate had effectively petered out because it had already been resolved.

However, as is well known, following the collapse of the Soviet Union, liberal economists brought these arguments—which had been refuted in the Mises School’s time—back into the spotlight with even greater vigor, this time presenting the damage inflicted on socialism by modern revisionism as irrefutable proof of their claims. Yet the damage modern revisionism inflicted on the socialist economy was, among other things, precisely the result of the economic “reforms” that were based on the very theoretical foundations of the Mises School!

To avoid digressing from our topic, we will not delve here into the presuppositions upon which Mises and Hayek built their theses—and which Marx, for his part, often mocked (the invisible hand of the market, free-market mechanisms and actors, etc.)—but will simply note the following: First of all, their theories regarding the role of price in the face of “scattered information” are incompatible with monopolistic capitalism. Prices are no longer determined by the “free market.” The price dictates of monopolies are plain to see. The claim that “scattered information” in the free market rationally allocates resources is nothing more than a hollow assertion in the face of today’s realities: income disparities, gaps in housing and healthcare, crumbling infrastructure, and so on. As for planning, this is, so to speak, the other “invisible hand” of today’s monopolistic capitalism. And this is no coincidence, for as the scale of economic activity expands—concentrating and centralizing production within monopolies, and thereby effectively socializing production—planning becomes a necessity. Today, monopolies not only engage in large-scale planning within their own structures but also, as parent companies, set transfer prices for goods, services, and other matters among their affiliated companies on an international scale; moreover, these transfer prices are not adjusted according to current market prices but are instead determined based on factors such as efficiency, which function as planning elements.

The inherent contradiction of monopolistic capitalist planning lies in the fact that it is carried out despite—or, more accurately, precisely because of—the anarchy and competition stemming from the private ownership of the means of production and permeating the economy as a whole. Each monopoly, by planning its own actions within this environment of anarchy and competition that governs production as a whole—that is, by seeking only to protect itself from its destructive effects—actually exacerbates this anarchy. It delegates the task of overcoming the contradictions and inadequacies—which arise from the profit-driven nature of production and the fact that the exchange of goods is based on value rather than use value, and which cannot be resolved through individual planning—to their governments and central banks. It should be clear that such contradictions in planning would not exist in a socialist economy that abolishes private ownership of the means of production and does not produce for exchange value—that is, one in which interpersonal relations (and thus the social character of production) are not experienced through the prism of value and the market. For central planning under socialism is carried out not in spite of the social character of production, but rather as a necessity of that very character.

Undoubtedly, the primary challenges in the construction and development of a socialist economy are not technical. The working class and labourers must persist in their efforts to raise the level of consciousness and cultural development demanded of them by the new relations of production, bringing the productive forces to a level of development and efficiency compatible with these relations, and the continuous review and renewal of the organizational and political dimensions of the formulation, monitoring, and oversight of the political and economic decisions that serve this purpose—in light of the challenges arising from ongoing class struggles—are among the key aspects. However, as a social system free from the constraints of the capitalist mode of production in utilizing and evaluating advances in science and technology, socialism needs to advance science and technology—precisely in order to successfully navigate the issues mentioned in the previous sentence—and to benefit from them as a productive force. Digital and AI technologies, for example, offer immense new possibilities for the large-scale automation of production under socialism. They not only enhance the ability to overcome many of the challenges encountered in previous experiences with socialist planning and improve the accuracy of resource allocation but also dramatically expand the capacity to monitor changes in real time through digital feedback mechanisms. Undoubtedly, it would be speculative to detail today exactly what changes new technologies—particularly in the organizational forms of socialist planning—will make possible and necessary. However, it is clear that they offer new opportunities for resolving many issues in the construction and operation of a socialist economy, foremost among them the productivity of the productive forces.

From this perspective, and considering the new possibilities offered by today’s science and technology, it seems that the problem of “distributed information” (in their theses, the role of price was linked to the transmission of this information), on which Mises and Hayek based their claims that a socialist economy could not function, has been solved almost “spontaneously” as a technical problem by digitalization and AI technologies. Hayek based the superiority of capitalism on its ability to provide “distributed information” through the market mechanism and thus make the distribution and direction of resources and alternative investments efficient. However, for example, the digital and AI technologies used in Walmart and Amazon today provide all the information about what is needed and when, whether it be material, consumer, or supply chain requirements, directly and in seconds, without needing the mediated information of the market and in a much more efficient way. Moreover, Amazon anticipates demand and predicts orders by using massive data analytics and artificial intelligence.[^[23]]

In other words, technological advancement as an element of productive forces renders Hayek’s “disperse knowledge” thesis meaningless. So, if capitalism no longer holds this dominance, what need is there for private ownership of the market and the means of production? Given technologies that instantly transmit information about needs and change, is there any basis left, even theoretically, for defending the impossibility of planning social production?

Of course, a socialist society would not use today’s digital infrastructure and AI technologies, which are trained for profit maximization, control, and surveillance, as they are. The priorities and functions in the algorithms would be changed according to the priorities of social production and the needs of society. Moreover, new algorithms would be developed. Thus, these technologies, shaped by capitalist relations, would not have the destructive power they currently possess against labour and social life; on the contrary, they would be transformed into levers to increase the free time that workers and labourers need to improve themselves in many ways. Of course, the struggle against the way AI technologies are used and the resulting social, cultural, and societal problems cannot be left to a future socialist society. The breadth of the range of formations and uses of AI technologies naturally increases the effects arising from their capitalist use against labour, the lower classes of society, and the people. Struggles against these effects are increasing. Even as these lines are being written, news of various protests and actions by Google and Meta[^[24]] employees has emerged.[^[25]] Following Google’s agreement with the Pentagon, the company’s employees in the UK intensified their unionization efforts and demanded that the company recognize the communications union CWU and the Unite union as their joint representatives. They explained that this demand was made urgent by their struggle “against the use of Google’s AI models in violation of international law, such as in Israel’s genocide against the Palestinians.”[^[26]] Prior to this, there had been months-long struggles by Amazon workers in various countries.

These reactions and struggles also show that the monopoly created by new technology companies, especially the “magnificent seven” in the US, in data and communication infrastructure must be broken; AI technologies must be regulated with laws that prioritize the interests of the working people in research and applications; new legal regulations must be created regarding the control rights of workers and the public; and trade union struggle and organization must be developed against the attempts of capital to deskill the AI ​​labour force in workplaces and to intensify the yoke on labour, devaluing it and using it as a justification for dismissals. Of course, this struggle must also include organizing opposition to trade union lines that treat technology as if it were a law of nature brought down to earth by the god of competition.

***

Both digital technologies and AI as a special dimension among them, as well as genome sequencing and biotechnology or nanotechnology or renewable energy technologies, etc., in short, the developments in science and technology, not to mention those before, in just the last 30 years[^[27]], represent great strides in the advancement of the productivity of social labour. However, when we look at how much these advancements have benefited society as a whole, especially the workers, and moreover, how much their lives have improved, the picture falls far short of what could have been. The fundamental reason for this relatively limited, distorted, and particularly destructive reflection of the productivity of social labour on the improvement of social life and welfare is the relationship with capital and the practices and impositions necessitated by its existence.

This disproportion, which has grown in relation to the level of development of the productive forces, and its manifestations—such as ongoing hunger, poverty, increasing insecurity, long and intense working hours and stress, restrictions on rights and freedoms, wars and conflicts, the climate crisis, and environmental disasters—stem from the monopolistic bourgeoisie’s drive to maintain and perpetuate its dominance at all costs, a dominance whose sole aim is to increase its own accumulation of wealth. The level of productivity of social labour, which it has spurred on for this petty purpose, enriches it on the one hand. However, on the other hand, this enormous disproportion it creates for the overwhelming majority of society reveals that this exploitative class is the greatest obstacle to the humanization of all aspects of social life. While one aspect of this disproportion is the increase in destructive manifestations, the other aspect demonstrates not only the urgency of its solution but also how advanced the preconditions for a socialist society, which envisions the abolition of capitalist relations, have become. In this respect, digital technologies and AI should be used not as an opportunity to ask whether monopolistic capitalism improves the productivity of labour and, as a component of it, science and technology (since this is essential for its large-scale production process), but rather as an opportunity to ask why social and societal problems and contradictions continue to grow despite these developments.

1. Bloomberg HT (2026) “Samsung Reaches Historic Milestone: $1 Trillion,” https://www.bloomberght.com/samsung-tarihi-kilometre-tasina-ulasti-1-trilyon-dolar-3776769 ↑
2. Say, C. (2018) “50 Questions About Artificial Intelligence,” Science and Future Library, Istanbul. ↑
3. This definition, which appears in Jason Resnikoff’s article titled “Objecting to the Idea of Progress: Labour’s Test with Artificial Intelligence,” is attributed to Margaret A. Boden, a recognized authority in the field. See Yapay Zeka ve İşin Geleceği (Artificial Intelligence and the Future of Work – 2025), Edited by Arif Koşar, Kor Publications, p. 184. ↑
4. This reproduction is, of course, not merely a repetition. The fact that AI technologies operate at enormous computational speeds also makes it possible to make new discoveries in an extraordinarily short time. ↑
5. Bender, E. M. et al. (2021) “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?”, In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’21), Association for Computing Machinery, New York, NY, USA, 610–623. https://doi.org/10.1145/3442188.3445922 ↑
6. The prerequisite for Taylorism’s separation, standardization, and even pitting against one another of physical and mental labour—as Taylor himself noted—is the systematic documentation of the worker’s “traditional knowledge.” The system for separating labour into physical and mental categories was developed through the recording of this knowledge and experience. See the article titled “The Automation of Death from the Employer’s Perspective: Artificial Intelligence and the Expropriation of Mental Labour” in this file. ↑
7. This data is collected, cleaned, and standardized into a unified format by low-wage “clickworkers” in various African countries, India, and elsewhere; these tasks are a prerequisite for operations in massive data centers. ↑
8. For example, in the “manifesto” recently published by the Palantir monopoly—which provides “digital services” to the state apparatuses of various countries, primarily the U.S. military, intelligence agencies, and police, using its own proprietary algorithms, and is headed by the radical reactionary Peter Thiel—alongside other racist and reactionary ideas, it argued that AI and digital infrastructure must be placed directly under the control of the ruling class and the bureaucracy. ↑
9. https://t24.com.tr/bilim-teknoloji/yapay-zeka-bizi-aptallastiriyor-olabilir-mi,1317094?_t=1778676760119 ↑
10. Marx, K. (2011) Kapital: Cilt I, trans. M. Selik and N. Satlıgan, Yordam Kitap, Istanbul, p. 465 ↑
11. Marx, op. cit., p. 307 ↑
12. Marx, op. cit., p. 313 ↑
13. “… with every change in the productivity of labour, the necessary labour-time also changes. The productivity of labour is determined by a wide variety of conditions, including the average level of skill of the workers, the level of scientific development and its technological applicability, the social composition of the production process, the scope and efficiency of the means of production, and natural conditions.” (Marx, op. cit., p. 54) ↑
14. Marx, op. cit., p. 375 ↑
15. Marx, op. cit., p. 377 ↑
16. Marx, op. cit. ↑
17. Marx, op. cit., p. 392 ↑
18. Marx, op. cit., para. ↑
19. Marx, op. cit. ↑
20. Marx, op. cit. ↑
21. Schlaudt, O. (2021) Lenin, Castro, Bezos? The Idea of “Cybersocialism” in the Light of Historical Planning Debates, Dietz Berlin, p. 41 ↑
22. Dobb, M. (1973) The Economists and the Economy of Socialism, Suhrkamp Verlag, pp. 7–10 ↑
23. Using its patented “Anticipatory Shipping” technology, the company processes metrics such as consumers’ shopping history, click-through rates, and time spent on pages to pre-ship products to the warehouse closest to the customer. ↑
24. Meta employees distributed flyers at many of the company’s offices in the U.S. In their flyers, they objected to the installation of new software—which had just been rolled out and was designed to track employees’ computer usage habits to train AI models—on their computers, and called on colleagues to sign an online petition outlining their demands on this issue. ↑
25. According to a Bloomberg report, hundreds of AI researchers at Google wrote a letter addressed to CEO Sundar Pichai, stating: “We are Google employees, and we are deeply concerned about the ongoing negotiations between Google and the U.S. Department of Defense. As people working on AI, we know that these systems can centralize power and make mistakes.” This protest followed a dispute between the Pentagon and another major AI company, Anthropic. The dispute concerned the use of AI in military applications. The Pentagon wants to remove Anthropic and its “Claude” AI tool from the U.S. defense supply chain. Anthropic, meanwhile, announced in recent days that its new AI model, “Claude Mitos,” is too dangerous to be disclosed to the public, describing it as a new and powerful AI model that poses “unprecedented cybersecurity risks.” As a reminder, Anthropic refused to allow the unrestricted military use of its “Claude” AI model; the U.S. government, in turn, designated Anthropic as a “supply chain risk to national security” and “ordered federal officials to stop using Claude.” (see https://tr.investing.com/news/stock-market-news/google-calsanlar-pichaiden-askeri-yapay-zeka-isini-reddetmesini-istiyor-93CH-3869800) ↑
26. Cradle (2026) “Google AI workers unionize against tech giant’s role in US, Israeli war crimes,” https://thecradle.co/articles/google-ai-workers-unionize-against-tech-giants-role-in-us-israeli-war-crimes ↑
27. Not to mention the economic and political pressures exerted by monopoly capital regarding what should be researched in science and technology and in which areas efforts should be concentrated! ↑