Yang Tianyu: The Distribution Paradox of Capitalist Intelligentization and Its Alternatives
Currently, a new generation of artificial intelligence (AI) technology—characterized by deep learning, cross-domain integration, and autonomous control—is triggering profound economic and social transformations globally. On one hand, AI enhances production efficiency through technological innovation and creates new wealth; on the other hand, it exacerbates income polarization. This polarization manifests in several ways. AI leads to the substitution of machines for human labor, thereby reducing labor demand, suppressing wages, and causing a decline in the labor share of income—a phenomenon that can be termed the "capital bias" of AI. Compared to low-skilled labor, high-skilled labor can learn and utilize AI more quickly, thereby increasing its productivity at a faster rate. Consequently, the wage gap between high-skilled and low-skilled labor continues to widen, intensifying income inequality; this can be termed the "skill bias" of AI. Furthermore, AI tends to reduce jobs through automation in low-tech sectors while creating new positions in high-tech sectors, which also exacerbates income inequality—this can be termed the "task bias" of AI. Through these channels, AI technology leads to highly unequal patterns of income distribution.
From the perspective of Marxist political economy, phenomena such as the "capital bias," "skill bias," and "task bias" of AI technology do indeed exist in reality, but AI technology does not inevitably lead to them. Like the steam engine, electricity, and the computer before it, AI technology itself cannot bring about these biases. On the contrary, the means of production must first rely on specific relations of production before they can influence the relations of distribution. As Marx pointed out: "Machinery is no more an economic category than the ox which draws the plough is. The contemporary application of machinery is one of the relations of our present-day economic system, but the way in which machinery is utilized is something quite distinct from the machinery itself." "The modern factory, which is based on the application of machinery, is a social relation of production, an economic category." This means that the "way in which machinery is utilized"—or more specifically, the way AI technology is utilized—is what truly transforms AI into an economic category and impacts income distribution. The "way in which machinery is utilized" varies across different relations of production. That is to say, whether AI actually exacerbates income inequality through capital, skill, and task biases depends on the specific relations of production within which it is utilized and the manner of its utilization.
Scholars in Western economics have studied the income distribution effects of AI, but they generally treat AI technology itself as the primary explanatory variable, ignoring the distinction between AI technology itself and the social relations of production based upon it. Marxist political economy provides a unique perspective on this issue. Relations of distribution are essentially reflections of the relations of production; the impact of AI on income distribution cannot be divorced from the ownership structure under a specific social formation. In other words, "the structure of distribution is entirely determined by the structure of production; distribution itself is a product of production." If the modern factory based on the application of machinery is a social relation of production, then the modern factory based on the application of AI technology is a new type of relation of production, and its impact on income distribution will differ from that of traditional machinery. This new type of relation of production will manifest in different forms under different ownership structures of the means of production—such as socialist versus capitalist new relations of production—thereby exerting different influences on the relations of distribution (i.e., income distribution). Existing research lacks sufficient understanding of this point. To deepen our cognitive grasp of income distribution patterns in the AI era, we must analyze various new economic categories based on the principles of Marxist political economy. We must answer a series of questions concerning the new relations of production and distribution in the AI era, such as how changes in the mode of production determine these new relations, and what impact different forms of ownership have on income distribution.
The analysis in this article demonstrates that the new capitalist relations of production and distribution in the AI era still conform to the general law of capitalist accumulation from the era of large-scale industry, while also containing specific forms of internal contradictions. Conversely, the new socialist relations of production, as an alternative, offer the possibility of constructing a new type of distribution relationship that balances efficiency and equity in the age of intelligence.
I. The Capitalist Labor Process in the Era of AI
As an advanced form of the productive forces, AI technology first influences the mode of production, which in turn impacts the relations of production and distribution. The capitalist mode of production in the sense of Marxist political economy can be understood as the capitalist labor process. Therefore, this article first analyzes the impact of AI on the capitalist labor process—the typical manifestation of the mode of production in the world today.
1. The Reconstruction of the Labor Process by AI Technology
The historical evolution of the capitalist labor process is not a linear result of technological progress, but a product of the interaction between the logic of capital and technological conditions. From the spatial concentration of manual labor in manufacture [1] to the mechanized labor organization of the Fordist assembly line, and then to the flexible production systems under post-Fordist conditions, every "reconstruction of the labor process" reflects capital's updating of labor control methods in pursuit of the maximization of surplus value. Currently, AI technology has triggered a new round of labor process reconstruction. Its fundamental characteristic is not simple "automation" but "intelligentization"—namely, using brain-like algorithms, deep learning, and large model systems as the core to construct intelligent means of labor that can learn, adapt, and make decisions, which are comprehensively embedded in production, organizational, and management processes.
Unlike the mechanization and automation of the 20th century, the embedding of AI is not a simple substitution for labor but a systemic reconstruction of the "cognitive dimension" of the labor process. Labor is no longer characterized by the human brain controlling tools to complete tasks, but by the collaboration between humans and intelligent systems. In some scenarios, the intelligent system even becomes the organizer of labor (e.g., intelligent production scheduling, automated quality inspection), while the worker is demoted to an executor, a collaborator, or a "maintainer of the algorithm."
The comprehensive embedding of AI into the labor process has given rise to an unprecedented strengthening of the domination of dead labor over living labor, or the control of capital over labor. Some scholars argue that in human-machine collaboration, workers are responsible for setting goals, providing resources, and verifying results, while AI performs task decomposition, tool selection, and progress control, autonomously concluding the work once the goal is reached. In reality, the workers capable of "setting goals and providing resources" are merely a small percentage of high-skilled engineers, by no means the average worker. For the majority of laborers, compared to the modes of production prior to the intelligent era, they not only lack the autonomy to "set goals and provide resources," but are conversely increasingly controlled by intelligent machines. This control is reflected in the informatization, flexibilization, and fragmentation of the labor process.
From the perspective of informatization, the labor process in the intelligent era is highly dependent on a complete closed loop formed by information collection, data training, model iteration, and intelligent feedback. This ensures that workers no longer control the work tasks. Instead, tasks are "pushed" by intelligent platforms to specific workers for completion. From the perspective of flexibilization, AI systems can dynamically adjust the production rhythm, job sequence, and even the organizational structure of labor based on external variables such as order changes, market feedback, and supply chain disruptions. While this high degree of flexibility does improve production efficiency, it simultaneously exacerbates the instability and unpredictability of the labor process; labor becomes increasingly subject to "data changes" rather than the will of the laborer. From the perspective of fragmentation, intelligent labor platforms decompose complete labor tasks into multiple modules distributed to different workers. Each laborer completes only a local, standardized unit of labor and lacks awareness and control over the overall product flow, leading to a further erosion of labor subjectivity and the sense of meaning in work.
Of course, these trends of informatization, flexibilization, and fragmentation already existed in the post-Fordist era, but the emergence of intelligent machines makes these trends fundamentally different. The post-Fordist labor process required "multi-skilled workers" and "work teams comprised of representatives from R&D, production, and sales departments" to manage the production flow. To put it simply, it required experienced "multi-skilled workers" to command from a dispatch room. In the era of intelligence, however, this role has been replaced by intelligent machines. In other words, intelligent production systems driven by algorithms possess autonomous learning capabilities and can make dynamic decisions through real-time data analysis. Thus, they can replace "multi-skilled workers" with higher efficiency while further strengthening the domination and control over living labor.
2. The Formation Mechanism of "Capital Bias" in the AI Era
The concept of "capital bias" proposed by Western scholars can be understood as the tendency of firms to substitute capital factors for labor factors as capital investment increases, or to replace human labor with machines, resulting in a decline in the labor share of income. In the context of Marxist political economy, "capital bias" is essentially the decline in the ratio of necessary labor time to surplus labor time, or an increase in the rate of surplus value. AI indeed leads to "capital bias"—an increase in the rate of surplus value—but this is not an issue with AI technology itself; rather, it is endogenous to the capitalist labor process in the AI era.
The role of AI in increasing the rate of surplus value is first manifested in the growth of relative surplus value brought about by the improvement of labor productivity. AI technology significantly enhances output levels per unit of time through means such as process optimization, dynamic scheduling, and machine self-learning, which can bring extra surplus value to the firms adopting this technology. In pursuit of extra surplus value, all firms have an incentive to adopt AI technology. The cumulative result is a reduction in the value of the means of subsistence, or the value of labor-power, thereby reducing necessary labor time and expanding the share of surplus labor time.
In addition to this relatively intuitive effect, AI can also achieve an increase in absolute surplus value by adjusting labor time. Marx believed that the production of absolute surplus value manifests as the lengthening of the working day or the shortening of rest periods. In the AI era, the adjustment of labor time takes on more covert and precise forms. Specifically, AI systems can achieve precise scheduling of workers' labor time through algorithmic control, behavioral prediction, and automatic adjustments of labor intensity. For example, in an intelligent manufacturing plant, an operator's movements are recorded by sensors on an intelligent wristwatch and their efficiency is evaluated in real-time; the employee's workload and working hours are then dynamically adjusted based on the evaluation. Consequently, capital no longer needs to rely on human-controlled clock-in systems to extend labor time. Instead, through seemingly non-human means such as "system reminders" and "abnormal behavior warnings," it integrates the fluctuating working hours of employees—affected by energy, mood, habits, or physical condition—into continuous, high-efficiency working time. This is an increase in surplus value achieved by intensifying labor.
In the traditional capitalist labor process, the capitalist application of machinery "converts the whole of the laborer's time and that of his family into labor-time at capital's disposal for its own valorization." In the AI era, the capitalist application of intelligent machines can convert the entirety of a worker’s working hours into high-efficiency labor time at capital's disposal for valorization. This is a new type of "time exploitation" that likewise serves to increase the rate of surplus value.
The substitution of labor by AI also increases the rate of surplus value. The popularization and application of AI not only replace physical labor but have also begun to systematically erode the fields of intellectual and cognitive labor. The development of technologies such as generative AI, intelligent algorithms, and brain-like computing means that knowledge-based labor, service labor, and even creative labor are being progressively replaced by machines and algorithms. This substitution has eliminated a large number of jobs, while the new positions created by AI may not necessarily fill the gap. The American scholar [Jeremy] Rifkin found...
"These new professionals—the so-called symbolic analysts and knowledge workers—come from the worlds of science, engineering, management, consulting, education, marketing, media, and entertainment. Although their numbers will continue to grow, their quantity is truly negligible compared to the number of workers replaced by the new generation of 'thinking machines.'" Unlike the steam engine, electricity, and computer technology, Artificial Intelligence (AI) technology inherently possesses the characteristic of substituting for human labor. Although there is currently no obvious evidence to prove that AI has indeed substantially reduced total job opportunities overall, there is no doubt that it will have an impact on the labor market. In this context, wage laborers have intensified their competition for job opportunities, producing two consequences: first, laborers are forced to accept longer working days and higher labor intensity; and second, the value of labor power is suppressed. Clearly, both of these consequences serve to increase the rate of surplus value.
II. Labor-Capital Relations and Class Structure in the AI Era
The capitalist labor process in the AI era will inevitably affect the relations of production and relations of distribution. In Marxist political economy, the ownership of the means of production is the decisive element in the relations of production; it determines the mode of distribution, as well as the status and mutual relations of individuals. Under capitalist ownership, AI relies on the capitalist ownership of the means of production to exert a profound influence on labor-capital relations and class structure, thereby determining the relations of distribution.
1. Labor-Capital Relations in the AI Era
In traditional capitalist relations of production, labor power becomes a commodity, and the employment relationship serves as the institutional basis for extracting surplus value. Entering the AI era, this basic framework has not undergone fundamental change, but AI technology has greatly altered the specific manifestations of labor-capital relations.
First, the employment relationship exhibits a trend toward "de-physicalization." Unlike traditional "embodied" labor centered on the factory, the labor process driven by AI is increasingly stripped away from physical space. Relying on cloud computing, remote collaboration, and automated decision-making systems, capital no longer needs labor to be concentrated in a specific location; instead, it exercises instantaneous command and remote management over dispersed individuals through intelligent platforms. This development of living labor becoming "off-site" essentially strengthens the real subsumption of labor under capital while simultaneously depriving laborers of the social space for face-to-face consultation and collective struggle. Second, the employment relationship exhibits a trend toward "fragmentation." Through technical means such as task modularization, intelligent scheduling, and real-time evaluation of individual performance, AI decomposes the complete labor process into "micro-labor," "instantaneous labor," and "on-demand labor," greatly enhancing capital's ability for the flexible deployment of labor power. This labor process not only reduces the autonomy of the laborer but also forms de facto "non-employment" labor—that is, the labor itself still exists and capital's command over labor exists, but the continuous employment relationship seems to have vanished. Consequently, systems such as traditional labor law protections and collective bargaining, which safeguard the rights and interests of laborers, become difficult to apply. From a historical perspective, this trend is not a dissolution of the employment system but rather an advanced stage of it—one where capital replaces personal control with technical control. While the shell of labor-capital relations is fragmented, the essence of the exploitative relationship becomes increasingly hidden and solidified.
Under the dominance of AI, another major change in labor-capital relations is the new asymmetric relationship in bargaining power between labor and capital. The capitalist side uses the advantages of intelligent technology to strengthen its control over labor, while laborers lose their ability to engage in effective bargaining due to the organizational and informational dilemmas of the AI era. On the one hand, the bargaining power of the capitalist side is unprecedentedly enhanced. Intelligent platforms can obtain labor performance in real-time, predict behavior, and optimize labor allocation, which endows capital with a "managerial power" that surpasses any previous era. On the other hand, the bargaining power of laborers continues to weaken. The "off-site" nature of living labor and the "fragmentation" of employment relationships brought by AI technology make it difficult for laborers to organize collective action. Mechanisms in traditional labor-capital relations, such as trade unions, collective bargaining, and legal rights protection, become ineffective or marginalized under the intelligent labor system. As a result, individual laborers can only accept the arrangements of intelligent algorithms, and their market value is instantly evaluated and priced by those algorithms. In this way, the roles of professional development, skill accumulation, and job seniority in the traditional sense are all seriously undermined. All this implies that the bargaining power of laborers has dropped significantly compared to the traditional industrial era.
It is worth noting that this new asymmetric relationship does not stem solely from the technological gap; more crucially, it stems from the inequality in the structure of ownership. That is to say, the capitalist private ownership of the means of production allows AI technology to be monopolized, enclosed, and privatized by capital, making it an extension of power rather than public wealth. This structural problem determines that even if laborers master some skills, it remains difficult to reverse the pattern of bargaining imbalance.
2. "Skill Bias," "Task Bias," and New Mutations in Class Structure
AI not only changes labor-capital relations but also reshapes the social class structure, manifesting primarily as "skill bias" and "task bias." Driven by AI, the labor market is polarized into two extremes: at one end are high-skill laborers engaged in non-routine complex labor (such as technology development and algorithm optimization); at the other end are low-skill laborers engaged in non-routine simple labor (such as data labeling and logistics delivery). Medium-skill laborers are easily replaced by AI because they engage in programmed, routine work, leading to a continuous decrease in their employment proportion. This can be understood as "skill bias" in the AI era. Furthermore, AI has created a large number of new jobs in related industries, mainly concentrated in fields such as algorithms, machine learning, intelligent chips, and robotics—for example, data scientists, machine learning engineers, and software architects. However, at the same time, a considerable portion of employment positions in traditional sectors face the risk of being replaced by AI. This can be understood as "task bias" in the AI era.
Under the interaction of "skill bias" and "task bias," two new types of laboring classes have appeared in capitalist society: the emerging labor aristocracy and the digital proletariat. The so-called emerging labor aristocracy refers to those high-skill laborers who master key AI technologies, data resources, and innovation capabilities, and are able to participate in the design and control of the intelligent production process. Although this group remains controlled by capital, they obtain higher salaries, benefits, and social status due to their "irreplaceability" for capital valorization. Capital even partially absorbs this group as allies of the technological bourgeoisie through methods such as equity incentives and platform partnership systems. The digital proletariat includes two types of people: first, laborers in traditional industries who are directly replaced or marginalized by AI, such as those in low-skill manufacturing, basic services, and traditional clerical work; and second, "digital laborers" who are controlled by algorithms, face extreme income volatility, and lack social security and collective action capabilities, including flexible employees, online crowdsourcing workers, and short-term project workers. The digital proletariat lacks both material wealth and bargaining power regarding the value of their labor power. This new class structure means that in the AI era, not only will the gap between the laboring class and the capitalist class widen, but the income gap within the laboring class will also expand.
In the AI era, digital capital possessing advanced AI technology exhibits globalized and monopolistic characteristics, which will intensify "skill bias" and "task bias" and further lead to the expansion of income gaps within the laboring class. On the one hand, digital capital in developed countries can use digital platforms to promote the global expansion of employment relationships, outsourcing tedious tasks such as data labeling and content moderation to developing countries at extremely low costs, giving rise to a massive overseas reserve of cheap digital labor power. These workplaces, dubbed "AI sweatshops," pay meager wages as low as 30 cents for four hours of work. This move not only enables digital capital to capture surplus profits but also masks the direct relationship of capital's exploitation of labor within developed countries. Through this spatial layout, capital transfers part of the class contradiction to the international level, while simultaneously reducing the bargaining power of low-skill laborers in developed countries due to job outflows, leading to lower incomes or even unemployment. This invisibly intensifies "skill bias" and widens the internal income gap of the laboring class. On the other hand, by virtue of its technological monopoly and platform advantages, digital capital breaks the traditional process of the equalization of the rate of profit. Industrial capital in traditional sectors can usually only obtain the average rate of profit, while digital capital, by controlling platforms, algorithms, and data, can achieve high profit rates that far exceed those of traditional industrial capital. This has led to the formation of two different types of capitalist production within developed countries: one is industrial capital following traditional distribution logic, and the other is digital capital enjoying monopoly premiums. This differentiation suppresses the room for wage increases for laborers in traditional industries while enhancing the ability of digital capital to co-opt high-skill laborers. This intensifies "task bias," causing laborers in traditional industries to face the pressure of income stagnation or even decline, while a few high-skill laborers in the digital industry who master AI technology can obtain high compensation. The final result is the expansion of the income gap within the laboring class.
Some proposals put forward by Western scholars to solve "skill bias" and "task bias" through the rational utilization of AI technology cannot achieve the desired effects in reality. For example, some scholars believe that "skill bias" can be regulated by the supply of labor. That is, although AI increases the demand for high-skill labor, if the education system can increase the supply of high-skill labor at a faster rate, the contradiction between supply and demand will be alleviated and the growth rate of high-skill labor wages will slow down. This would reduce the skill premium and thus solve "skill bias." The problem with this proposal is that if "skill bias" were indeed alleviated in this way, its side effect would be the intensification of "capital bias." This is because it means the wages of all labor would be suppressed at a low level, and none would share in the growth dividend brought by AI. This is equivalent to reducing the ratio of necessary labor time to surplus labor time and increasing the rate of surplus value.
Another proposal is that AI can create new jobs for low-skill labor. Certain "non-routine tasks" that cannot be processed by computer programs, such as emotional interaction, teamwork, and other tasks requiring "soft skills," could bring job opportunities to low-skill labor. In reality, such industries have absorbed a large amount of labor employment; "all-media operators" [2] is one such profession. The employment directions of this profession include live streaming, short video creation, and new media operations. It cannot be denied that there are indeed many practitioners—such as network anchors, video bloggers, and professional mobile gamers—prompted by AI in reality, but these professions, like all small-scale producers, exhibit a trend toward polarization. In fact, they cannot reverse the expansion of the income gap brought about by "task bias."
III. The Distribution Paradox of Capitalist Intelligentization
In his analysis of large-scale machine industry, Marx once pointed out: "the machine in itself increases the wealth of the producers, while its capitalist application turns the producers into paupers in need of relief." In the AI era, the capitalist application of AI has produced similar results. It is undeniable that AI technology has indeed increased social wealth. However, under the effect of "capital bias," "skill bias," and "task bias" inherent in the capitalist labor process, laborers are not only unable to share in the newly added social wealth but must also accept the result of a decreasing share of labor income; most of them will even fall into the digital proletariat, lacking bargaining power. Unsurprisingly, AI has intensified income and wealth inequality in almost all developed capitalist countries. Moreover, the reconstruction of the capitalist labor process by AI technology has caused signs of failure in those institutions and policies in traditional capitalist society that safeguard the rights and interests of laborers. The original intention of these institutions and policies was to allow laborers to share in economic growth to a certain extent, but the capitalist application of AI makes these institutions and policies run counter to their original intentions, producing many paradoxes.
1. The Social Security Paradox: The Contradiction Between Flexible Employment and the Reproduction of Labor Power
The original intention of social security is to relieve laborers of anxiety and fear about the future, which is conducive to the reproduction of labor power. But in the AI era, the trends toward the informationization, flexibilization, and fragmentation of the capitalist labor process have caused the original social security system to exhibit phenomena of failure and collapse.
The impact of artificial intelligence (AI) on the social security system is first manifested in the weakening of the stability of labor relations. Stable labor relations are the prerequisite for workers to fight for their rights and organize collective actions. On this premise, mechanisms such as stable labor contracts in the traditional sense, trade union organizations, and collective bargaining can function effectively. In the era of AI, however, the informationalized, flexible, and fragmented labor process causes labor itself to be deconstructed into a form of "micro-tasks and instant feedback." Under these circumstances, capital does not need to employ workers on a continuous basis, resulting in a large number of flexible employees [3]. To pursue extraordinary surplus value, capital in the intelligent era increasingly adopts AI technology. The consequence is an increasing number of flexible employees, increasingly unstable labor relations, and increasingly fragmented worker organizations; trade unions and collective bargaining have thus lost their foothold. Evidence suggests that trade union organizations in developed capitalist countries have already been hit hard by the sharp increase in flexible employment and are at risk of becoming unsustainable. Deprived of the protection offered by unions and collective bargaining, the rights and interests of flexible employees lack institutional and organizational safeguards, and the level of social security they can obtain is consequently lower. Furthermore, flexible employees must pay their own social insurance premiums, yet it remains unclear who should bear the employer’s portion of the contribution. Requiring them to bear both the employer and employee contributions is unrealistic due to their generally low incomes, making it difficult for them to integrate into the existing social security system.
The impact of AI on the social security system is also reflected in the risk-shifting behavior of capital. By means of AI technology, capital has adopted employment methods such as flexible hiring, task-based dispatching, and remote collaboration, effectively outsourcing production and operational risks. While effectively evading traditional employer responsibilities, capital transfers increasing uncertainty to individual workers. Workers are forced to bear the continuous impact of unstable income, uncontrolled labor intensity, health risks, and technological turnover. This makes workers more in need of social security than in the traditional industrial era; however, unstable labor relations make it difficult for them to enjoy the labor protections associated with formal employment, forcing them to accept exploitation by capital.
A more profound issue is that the disintegration of this social security system not only leads to hardship for individual workers and a loss of social security, but also seriously erodes the system of social reproduction. Capitalist reproduction requires not only the reproduction of material means but also the reproduction of labor power. However, in developed capitalist countries, flexible employment groups who lack social security often postpone starting families due to widespread concerns about career prospects, and many even abandon childbearing altogether. Even those seemingly stable, non-flexible groups exhaust vast amounts of energy on career maintenance for fear of losing their established status, thereby delaying marriage and childbirth plans. The result is a decline in the birth rate across society. In other words, the collapse of the livelihood security system in the AI era actually destroys the social conditions for the reproduction of labor power. Moreover, we cannot expect to improve this situation through reforms of the social security system. Such reforms would require capital to reassume the social security responsibilities of an employer—that is, to increase variable capital expenditures in the form of social security contributions—which contradicts the logic of capital accumulation. Therefore, the "rational" behavior of capital will be as described: "When the bourgeoisie is faced with a tax increase, they always seek compensation by lowering wages or raising prices."
2. The Working Day Paradox: The Contradiction Between Job Opportunities and the Rate of Surplus Value
Many Western scholars have realized that the popularization of AI technology will reduce job opportunities. Consequently, they have proposed a once-effective strategy: shortening the working day. Every previous industrial revolution significantly reduced job opportunities, and the ultimate solution was not merely that new technologies created new jobs, but also that shortening the working day played a role that cannot be ignored. For example, the first industrial revolution in the 19th century brought a significant increase in labor productivity, and working hours were reduced from 80 hours per week to 60. The second industrial revolution in the 20th century similarly saw a steady increase in labor productivity, and working hours were further reduced from 60 to 40 per week. The effect of this strategy should not be underestimated. In today's AI era, if working hours could be reduced from 40 to 30 or even 20 hours per week, it would be equivalent to doubling the original job opportunities, which would clearly alleviate the impact of AI technology on employment.
While this idea is inherently good and has worked historically, it may struggle to be effective in the AI era. The main problem with shortening the working day is its contradiction with the capitalist's goal of pursuing extraordinary surplus value. Obviously, shortening the working day cannot come at the cost of a proportional reduction in wages; otherwise, the policy will not alleviate "capital bias" to any extent and will lose its practical significance. Therefore, plans to shorten the working day in developed capitalist countries involve a significant reduction in labor time accompanied by a smaller reduction in wages. For example, a French proposal suggested reducing the work week from 39 hours to 33 hours while reducing wages by 5%, which could increase employment by 10% and create 2 million new jobs. However, this is equivalent to a substantial reduction in surplus labor time while necessary labor time only falls slightly; the result would be a sharp drop in the rate of surplus value. This contradicts the interests of capital. To date, this plan has not been implemented.
Another problem with this vision is that the popularization of AI may not only fail to shorten the working day but may actually extend it. After introducing AI, enterprises can lay off large numbers of workers, creating an industrial reserve army [4] of the unemployed. Because they face competitive pressure from this industrial reserve army, those who remain employed are forced to work longer hours. Enterprises tend to use a small amount of labor to work longer hours rather than using more labor for less work; this is "rational" behavior. By doing so, enterprises can save on additional fringe benefit expenditures such as health insurance and pensions, while paying only half-pay for overtime work.
The deeper issue is that the shortening of the working day during historical industrial revolutions did not occur because capitalists found their conscience. On the contrary, it was the result of continuous struggle by the working class. However, the informationalized, flexible, and fragmented labor process of the AI era has turned a large number of workers into flexible employees, thereby weakening the power of trade unions. Under these conditions, the pressure on capital is actually reduced; thus, the motivation for capital to shorten the working day will weaken rather than strengthen. Therefore, this strategy, which was effective in the past, will encounter more contradictions in the AI era, and the difficulty of implementation has actually increased.
3. The Basic Income Paradox: The Contradiction Between Basic Income and the Rate of Surplus Value
Basic income is another strategy that Western scholars hold high hopes for in dealing with AI. Basic income refers to a cash subsidy regularly distributed by the government to all citizens or residents to guarantee basic living needs. Its core characteristic is that the state provides unconditional and universal transfer payments to all citizens. Logically, basic income can not only provide a basic safety net for the unemployed in the AI era, relieving them of their worries, but also motivate them to engage in more creative work, driving social vitality and innovation.
This vision is also beautiful and has been implemented on a small scale in developed capitalist countries. However, it faces a fundamental problem: what is the source of basic income? If basic income is only distributed in small-scale pilots, the government can manage by relying on public spending without changing the current tax structure or social security contribution amounts. The problem is that if AI causes large-scale unemployment, the amount and scope of basic income implementation would expand significantly, forcing the government to raise tax rates. This creates two problems. First, as AI technology is first applied in developed capitalist countries, basic income policies will inevitably be implemented in these countries first, while developing countries will not implement them for the time being. Thus, developed capitalist countries raising tax rates to fund basic income will inevitably impact the international competitiveness of their domestic enterprises, which contradicts the goal of domestic capital to pursue a higher rate of surplus value. Second, the increase in tax rates will lower the rate of surplus value. To achieve their goal of a higher rate of surplus value, enterprises will inevitably adopt more AI technology, resulting in more unemployed people needing basic income subsidies. This, in turn, forces the government to further raise tax rates, forming a self-reinforcing vicious cycle.
Furthermore, it is doubtful whether basic income will motivate the unemployed to "engage in more creative work." Unlike unemployment benefits that have a fixed period of receipt, basic income is not canceled because the recipient finds a job and essentially has no time limit. This form of distribution is actually more likely to "cultivate idleness" rather than inspire creative work. The American scholar [Jeremy] Rifkin suggested that basic income could be linked to community work—that is, the unemployed must provide a certain amount of labor in non-profit sectors such as communities to be eligible for basic income. If such a practice were actually implemented, the unemployed might no longer be idle, but at the same time, they would be participating in community work out of economic necessity. This essentially differs little from traditional public works projects and has nothing to do with "more creative work."
IV. Institutional Advantages of Socialist Intelligentization
The distribution paradox of capitalist intelligentization is endogenous to the capitalist labor process and its derived capitalist relations of production and distribution. However, this is not the only possible outcome in the era of AI. From the perspective of Marxist political economy, the root of the distribution paradox of capitalist intelligentization lies not in AI technology itself, but in the fact that AI technology is privatized by capital and alienated into a tool for extracting surplus value. The fundamental way to resolve the distribution paradox of capitalist intelligentization lies in breaking through private ownership and the logic of capital to find a rational alternative. Therefore, it is necessary to delve into how socialist intelligentization affects income distribution.
1. Compatibility of AI with Socialist Relations of Production
AI technology itself does not differ across different relations of production, but its impact on income distribution will be vastly different. Under the capitalist system, AI acts as an advanced productive force whose development is confined within the profit drive and technological alienation of the logic of capital, thus failing to release its proper social value. The socialist system, through public ownership and people-centered distribution mechanisms, possesses the capacity to embed AI into the New Development Philosophy centered on the people, thereby constructing a new model of a "technological community" and "value sharing."
From a purely technical standpoint, the characteristics of informationalization, flexibility, and fragmentation in the capitalist labor process also inevitably exist in socialist China. However, this does not mean that "capital bias," "skill bias," and "task bias" will emerge under the socialist system in the same way. The reason capitalist intelligentization cannot solve these problems is that the distribution of the means of production determines the distribution of products. The capitalist ownership structure adapted to the capitalist labor process ensures that any solution to these "bias" problems contradicts the capitalist's goal of pursuing a higher rate of surplus value, making these problems unsolvable. Conversely, in socialist relations of production, public ownership holds the primary position and the state-owned economy plays the leading role, which is a very significant institutional advantage. Under this system, all the people of a country own a vast amount of state-owned assets, which makes the distribution of wealth and income in that country entirely different from that of capitalist countries. As the French economist [Thomas] Piketty and others have noted, the high share of public ownership in China’s national wealth has mitigated the intensification of income inequality; this also explains why the growth of income inequality in China is lower than in the United States (where the share of public wealth is negative).
Vast state-owned assets provide a solid foundation for all people to share the dividends of AI technology. Whether it is basic income, universal intelligence training, or digital infrastructure popularized in urban and rural areas, all can receive support from the state-owned economy. Thus, those issues that plague developed capitalist countries...
Questions such as "capital bias," "skill bias," and "task bias" find a readily available path toward resolution. Taking "capital bias" as an example, socialist intellectualization similarly increases labor productivity, adjusts labor time, and replaces human labor. The state-owned economy will consequently obtain greater profits; however, because the ownership of the state-owned economy belongs to the whole people, these profits can logically be transformed into a basic income shared by all citizens through reasonable policy and institutional settings. The basic income policy, which is difficult to truly implement in developed capitalist countries, faces no institutional obstacles in socialist China. Thus, through the redistribution of profits, the state-owned economy can compensate for the loss of labor’s share of income in primary distribution. Furthermore, this compensation is shared by the whole people rather than exclusively by employees of state-owned enterprises (SOEs), thereby opening a path to resolving "capital bias." Taking "skill bias" and "task bias" as further examples, socialist intellectualization also creates gaps between high-skilled and low-skilled workers, and between emerging and traditional sectors; yet the state can provide reasonable income to low-skilled workers and traditional sector employees replaced by artificial intelligence (AI) through the distribution of basic income, thereby greatly alleviating these problems. In this sense, socialist relations of production possess a quite strong compatibility with AI technology.
2. Common Prosperity: The Key to Cracking the Distribution Paradox of Capitalist Intellectualization
The concept of "universal basic income" proposed by Western scholars is closer in meaning to "guaranteeing basic living needs" and does not contain the concept of common prosperity. Socialist countries find it easier to implement a true basic income policy, but basic income here is not limited to guaranteeing the basic living needs of the whole people. The essential requirement of socialism is common prosperity; therefore, the solution for income distribution problems in the process of socialist intellectualization should also take common prosperity as its goal. Capitalist intellectualization not only causes "capital bias," "skill bias," and "task bias," but also makes it difficult to resolve the worsening distribution caused by these biases due to the existence of the social security paradox, the working day paradox, and the basic income paradox. Conversely, socialist intellectualization aimed at common prosperity not only ensures the compatibility of AI and socialism regarding the basic income issue but also helps to comprehensively crack the distribution paradoxes of capitalist intellectualization.
In China, the superiority of socialist intellectualization lies in the fact that, through the leadership of the Communist Party of China (CPC) and the strategic intervention of the state governance system, it can provide a systematic solution to crack the distribution paradoxes of capitalist intellectualization. The Party's leadership ensures macro-strategic determination, allowing the path of technological evolution to remain anchored to the goal of common prosperity rather than simply serving the logic of capital or the local interests of technical oligarchs. Through top-level design, the Party's leadership embeds the principles of fairness and justice into the entire process of AI technology application and effectively guides the development direction of the intelligent economy through the state governance system, making it serve the overall goal of common prosperity. This governance mode, which uses institutional rationality to harness technological rationality, not only overcomes the limitations of spontaneous market regulation but also lays a solid foundation for constructing a more just distribution order on the basis of highly developed productive forces.
Socialist intellectualization must first crack the social security paradox. An increase in flexible employment is an inevitable result of the popularization of AI technology; however, this does not mean that social security for flexible employees will necessarily collapse. The socialist system has multiple solutions for the social security of flexible employees. The most fundamental solution still stems from the support of the state-owned economy. Since the owners of state capital are the whole people, all laborers can share in the economic growth of the AI era by enjoying the "capital returns" of state capital. Consequently, the basic income provided by the state is not merely as simple as "direct cash transfers," but can also be used to supplement the social security contributions of flexible employees. Currently, the Chinese government has not only issued policies to supplement social security funds with returns from state capital, but nearly 50% of the expenditures in the state capital operations budget have been transferred to the general public budget for social security and livelihood expenditures. Flexible employees brought about by AI technology, as part of the whole people, naturally enjoy this policy dividend. Moreover, if this policy can be appropriately expanded to include objectives such as increasing the participation rate and contribution base for flexible employees and making up for gaps in their social security payments, it would provide an effective way to solve the social security paradox. Furthermore, under the socialist system, mechanisms such as stable labor contracts, labor unions, and collective wage consultation [5] will not weaken or disappear due to the popularization of AI technology. Under the leadership of the CPC, not only SOEs but also private enterprises must establish labor unions and collective wage consultation mechanisms, including emerging industries that extensively apply AI. For example, the All-China Federation of Trade Unions (ACFTU) has promoted the establishment of wage consultation mechanisms between 12 leading digital platform enterprises and unions/worker representatives, covering 17.8 million workers. Under these constraints, even if capital-owners subjectively wish to transfer risks to laborers through AI technology, they are objectively unable to do so. This fundamentally cuts off the path by which the application of AI technology leads to the disintegration of the social security system.
Socialist intellectualization also has the capacity to crack the working day paradox. This paradox arises because shortening the working day contradicts capital’s goal of pursuing a higher rate of surplus value. In a socialist market economy, enterprises also have requirements for pursuing profit rates, but this does not reach the point of making the shortening of the working day a paradox as it does under the capitalist system. The leadership of the CPC is a defining feature of China's socialist market economy system. If it becomes necessary to shorten the working day to alleviate unemployment caused by AI applications, the Party’s leadership is fully capable of requiring all enterprises of various ownership types to implement the shortening of the working day. Therefore, the "rational" behavior of certain firms in capitalist countries to extend the working day is unfeasible in a socialist country.
Regarding the basic income paradox in capitalist countries, socialist countries also have the ability to crack it. The root of this paradox lies in the fact that the scale of the state-owned economy in developed capitalist countries is too small, while they simultaneously carry huge fiscal deficits and national debt; therefore, the source of basic income can only be a substantial increase in tax rates, which will certainly contradict capital's pursuit of a higher rate of surplus value. In socialist China under the leadership of the CPC, since basic income can be sourced from a powerful state-owned economy, the problem of financing basic income can be solved with only a limited increase in tax rates. Thus, raising basic income primarily affects the distribution of state capital returns during the redistribution phase, and the tax burden on enterprises of all ownership types in the primary distribution phase will be greatly lightened. This cracks the contradiction between basic income and the rate of surplus value in capitalist countries. The role of basic income in the process of socialist intellectualization does not stop there. In capitalist countries, even if basic income is distributed to all citizens, it can only serve as a bottom-line safety net and is insufficient to make all citizens prosperous based on basic income alone. In a socialist country, the state-owned economy, on the basis of sharing the economic growth of the AI era, can distribute its capital returns to all citizens in the form of basic income until common prosperity is achieved. The key difference between the two is that the essence of basic income in capitalist countries is "social welfare," which allows ordinary citizens only to attain basic subsistence (wenbao [6]) but not to fully share in the economic growth of the AI era; whereas the essence of basic income in socialist countries is "capital returns," through which ordinary citizens can fully share in the economic growth of the AI era. Clearly, the root of this difference lies in the difference of systems, or what can be termed the difference in the relations of production.
3. Promoting All-Round Human Development: The Ultimate Goal of Socialist Intellectualization
General Secretary Xi Jinping once pointed out: "Promoting common prosperity and promoting all-round human development are highly unified." In fact, promoting all-round human development is not only the goal of common prosperity in a socialist society but also one of the goals of communism. Marx pointed out in Capital that communism is "a higher form of society, a society in which the full and free development of every individual forms the ruling principle." To achieve the full and free development of the human being, it is necessary for individuals, after escaping the shackles of alienated labor and private property, to fully develop their physical strength, intelligence, talents, and personality. As early as the 1980s, Chinese scholars already pointed out that under conditions of high automation brought by AI, when those "material production processes" are replaced by intelligent machines, humans can break free from the constraints of the machine and finally achieve the abolition of the division of labor. In the classic works of Marxism, the abolition of the division of labor precisely implies the full and free development of the human being. For example, Engels pointed out in Anti-Dühring that only by abolishing the "old division of labor" can "every individual be given the opportunity to develop and exercise all their faculties, physical and mental, in all directions; in which, therefore, productive labor will become a pleasure instead of a burden." It is evident that AI can logically promote the full and free development of humans. However, we must add a qualifier to this conclusion: only socialist intellectualization can promote full and free human development, whereas capitalist intellectualization does not have this effect.
The main problem with capitalist intellectualization is that, on the surface, it seems it can liberate workers from the shackles of machines; but thereafter, although the liberated workers possess a large amount of leisure time, they are likely not to use this time to fully develop their physical strength, intelligence, talents, and personality—that is, to engage in more creative work—but instead become idle. This is because capitalist intellectualization finds it difficult to solve the problem of alienated labor. In the Economic and Philosophic Manuscripts of 1844, Marx provided three definitions of alienated labor. First is the alienation of the worker from the product of his labor: "the appropriation of the object appears as estrangement to such an extent that the more objects the worker produces the less he can possess and the more he falls under the sway of his product, capital." "It is clear that the more the worker spends himself, the more powerful becomes the alien world of objects which he creates over and against himself, the poorer he himself—his inner world—becomes, the less belongs to him as his own." This obviously unjust distribution reduces the worker's enthusiasm for labor, leading to the alienation of the worker from the process of labor: "the external character of labor for the worker appears in the fact that it is not his own, but someone else’s... as soon as no physical or other compulsion exists, labor is shunned like the plague." All this is because the worker "looks upon this activity as service to another, under the control, the compulsion, and the yoke of another"—that is, the alienation of man from man, where the "other" refers to the capitalist.
Because capitalist ownership has not changed, the three levels of alienated labor mentioned above will not disappear even in the AI era. In capitalist society, even if a worker "liberated from the shackles of the machine" has the desire to fully develop his physical strength, intelligence, talents, and personality to engage in creative work, his necessary activities such as learning, training, entrepreneurship, management, and sales must all seek support from venture capital in the capital market, and most of his creative results must be surrendered to the owners of capital. This means his activities remain "under the control of another"; his "creative work" will still make the power of capital stronger, while very little belongs to him. Therefore, he will still view "creative work" as "merely a means to satisfy needs-the need to maintain physical existence" and will still "shun labor like the plague." Thus, we can understand why capitalist intellectualization is more likely to turn those unemployed people receiving basic income into idle people, rather than people engaged in "more creative work." As a result, in capitalist society, only a few will enjoy the economic growth brought by intellectualization, while the majority can only obtain basic income in the nature of social welfare, making the problem of income distribution impossible to resolve.
The Sinicization of AI follows a completely different path. In the New Era of AI, the strong state-owned economy in a socialist country possesses the capacity to resolve the problem of labor alienation. For instance, a socialist state can rely on state-owned capital to establish a specialized fund dedicated to subsidizing more creative work by its citizens. A portion of the resulting returns would serve as state-owned capital gains and ultimately be transformed into a universal basic income for all people; another portion would belong to the citizens themselves as capital gains, structured as "citizen technical shares." In this way, any ordinary citizen engaging in more creative work during their leisure time can simultaneously obtain returns from both state-owned capital gains and private capital gains. From the perspective of the citizen as a laborer, their product of labor at this point becomes a "human existence" [7] rather than an alien power, which fundamentally eliminates the root of the alienation of the products of labor within capitalist relations of production and achieves a reconciliation between labor and its product. Of course, socialist AI development may have more than just this one path for solving the alienation of labor products. However, the aforementioned path demonstrates that a socialist state has the capacity to resolve the problem of labor alienation in the AI era. The resolution of this issue will inevitably and greatly enhance the initiative of laborers, ensuring that after they are liberated from the shackles of the machine, they possess not only sufficient time but also sufficient motivation to fully develop their physical strength, intelligence, talents, and personality, thereby achieving free and well-rounded development. If those laborers who have obtained basic income and leisure time all have the motivation to do so, then not only will wealth gush forth [8] across the whole of society, but they themselves will also receive generous returns, and the income gap between them and the high-skilled laborers who master key AI technologies will be greatly narrowed. At this point, there will no longer be a distinction between a "new labor aristocracy" and a "digital proletariat," but only occupational differences based on whether one directly participates in the AI production process. This means that laborers throughout society will have achieved free and well-rounded development; the income gap arising from the popularization of AI will, consequently, be resolved naturally. Such an ideal state can be regarded as the ultimate goal of the AI era. Although it will not arrive quickly, socialist AI development can provide an actionable practical blueprint for achieving this goal, building a bridge that connects theory with practice.
Conclusion
The rapid development and widespread application of AI technology in the early 21st century have brought about pressures of income distribution inequality while significantly improving productive efficiency. This pressure does not stem from AI technology itself, but is endogenous to the changes in the mode of production and relations of production induced by the development of AI technology. Some Western scholars have noted the pressures on income distribution brought by AI, but they focus solely on AI technology itself while ignoring the role of the mode of production and relations of production. This makes it difficult for them to answer a series of important theoretical questions: for example, whether and how AI's impact on labor-capital relations and its reshaping of the social class structure affect the state of income distribution across society; whether those institutions and policies in capitalist society that protect the rights and interests of laborers can solve the income distribution problems of the AI era; and whether an economic base and superstructure founded on public ownership can provide a better choice for solving income distribution problems. Starting from Marx's logic of totality and the theory of surplus value, this article explores the laws of income distribution in the AI era from the perspective of the mode of production and relations of production. The analysis shows that income distribution problems in the AI era cannot be blamed on AI technology itself, but are closely related to the nature of the relations of production. Within capitalist relations of production, capitalist AI development triggers problems such as "capital bias," "skill bias," and "task bias," leading to further intensification of income distribution inequality. Those policies that might protect laborers' rights have encountered a series of contradictions in capitalist society and are already struggling to be effective. Conversely, socialist AI development as an alternative—based on the economic base and superstructure of a socialist society—has opened a new path for solving income distribution problems in the AI era and provided an actionable practical scheme for achieving the free and well-rounded development of the individual. In other words, AI will not inevitably worsen income distribution; advanced relations of production provide the opportunity to solve this problem.
General Secretary Xi Jinping has pointed out: "The new round of scientific and technological revolution and industrial transformation has powerfully promoted economic development, but has also brought profound impacts on employment and income distribution, including some negative impacts, which need to be effectively responded to and resolved." The impact of AI technology on income distribution is precisely such a problem that requires an effective response and solution. To thoroughly solve this problem, one cannot be narrowly confined to AI technology itself, but must return to the fundamental pursuit of Marxism: the realization of the free and well-rounded development of the individual. The essence of the income distribution problem in the AI era does not lie in AI technology itself, but in whether social fairness and the free and well-rounded development of the individual can be promoted through rational institutional design. Specifically: first, promote the socialization of new types of productive forces related to AI to prevent technical dividends from being monopolized by capital groups. Second, establish a social security and income redistribution system for the AI era and promote the implementation of universal basic income to alleviate structural poverty and class solidification. Third, reshape development goals, shifting from a profit orientation to a social value orientation, so that AI truly serves the free and well-rounded development of the individual rather than being reduced to a tool of exploitation. The society of the future should not be a cold world of machines dominated by intelligent algorithms, but a harmonious society where technology empowers, institutions provide protection, and intelligence is directed toward the good. In this historical process, China's path of socialist AI development and the strategy of common prosperity will provide important inspiration for the world, and will contribute Chinese wisdom and Chinese solutions to the realization of technical justice and social fairness in a true sense.