The Generalization of Artificial Intelligence and New Propositions for the Labor Theory of Value

General Secretary Xi Jinping has stated that artificial intelligence is a key driving force behind the new round of technological revolution and industrial revolution, and that accelerating the development of a new generation of artificial intelligence is a strategic issue concerning whether China can seize the opportunities of this new round of technological and industrial revolution. Today, with the continuous introduction of large models, the pace of AI generalization is also accelerating. AI has moved from the machine replacement of labor by specialized AI to an era of replacing the laboring subject itself. This not only means that, in the current era of AI generalization, the efficiency of human production has been greatly improved, but also indicates that generalized AI will inevitably have a revolutionary impact on the production structure of human society, and will thereby fundamentally affect the logic of social operation. Therefore, with the support of big data, large models, and powerful computing power, AI, by roughly imitating the generality that emerges from human intelligence, continuously intervenes in human production and life, and raises the fundamental question of "AI replacing humans" today on a broader level than "machines replacing humans." The change in the situation of "machines replacing humans" not only fundamentally questions the meaning of human existence, but also raises important practical issues such as the subjectivity and creativity of AI's "quasi-labor," the confirmation of property rights and social distribution of wealth "generated" by AI, and the meaning and value of human labor. Faced with major social difficulties arising from AI generalization, such as "unmanned factories," "the useless class," [1]1 "The useless class" (无用阶级) is a term popularized by historian Yuval Noah Harari, referring to a hypothetical social stratum rendered economically and politically redundant by advanced AI and automation. "the de-materialization of labor objects," "the digitalization of labor forms," and "new forms of human alienation," we must analyze and answer important questions raised by AI generalization: the essence and value of labor, the high-level automation of labor and the social subjectivity of labor, and whether the "human-like labor" of general AI endows value or creates value. Therefore, we must confront the new propositions for the labor theory of value raised by AI generalization.

习近平总书记提出，人工智能是新一轮科技革命和产业革命的重要驱动力量，加快发展新一代人工智能是事关我国能否抓住新一轮科技革命和产业革命机遇的战略问题。今天，随着大模型的不断推出，人工智能通用化的脚步也在不断加快。人工智能已从专用人工智能对劳动的机器取代，进入对劳动主体取代的时代。这不仅意味着在人工智能通用化发展的今天，人类生产的效率得到了极大的提高，而且表明，通用化的人工智能必将对人类社会的生产格局产生革命性的影响，进而也会从根本上影响社会运行的逻辑。因此，人工智能在大数据、大模型、强算力的加持之下，通过大致上模仿人类智能涌现出的通用性，使其不断地介入人类生产与生活之中，并在比“机器换人”更广的层面上提出了今天“人工智能换人”的根本问题。“机器换人”情态的变化，不仅从根本上考问人类存在的意义，更在现实上提出人工智能“类劳动”的主体性与创造性、人工智能“产生”的财富确权和社会分配、人类劳动的意义与价值等重要问题。面对人工智能通用化所出现的“无人工厂”“无用阶级”“劳动对象去物化”“劳动形式数字化”“人的新异化”等重大社会难题，我们必须分析与回答人工智能通用化所提出的劳动本质与价值、劳动的高阶自动化与劳动的社会主体性、通用人工智能的“类人劳动”是赋予价值还是创造价值等重要问题。因此，我们必须正视人工智能通用化所提出的劳动价值论的新论题。

The "Generalization" of AI and the Value of Replacement Technology

人工智能的“通用化”与取代技术的价值

Whether it is the early conviction in general artificial intelligence at the inception of AI, or the confirmation today, brought by large models such as ChatGPT and Sora, that AI will inevitably move toward generality, both are a kind of proof of the meaning and value of "replacement technology." The generalization of AI is essentially a high-level automation of the "automatic system of machinery," [2]2 "Automatic system of machinery" (自动的机器体系) is a concept from Marx's *Grundrisse*, describing a stage where the means of labor is transformed into an automatic system, reducing the worker's role to that of overseer. an expression of the independence, human-likeness, and universality of replacement technology as manifested by machines.

无论是人工智能产生之初对通用人工智能的确信，还是今天由ChatGPT、Sora等大模型带来的人工智能必将迈向通用性的确证，都是对“取代技术”意义与价值的一种证明。人工智能的通用化，实质上是“自动的机器体系”的高阶自动化，是以机器表达出来的取代技术的独立性、类人性和广泛性。

First, in the process of upgrading replacement technology, AI generalization is replacing the general forms and processes of human labor, distinguishing the form in which labor creates value from that of traditional production and the automation era. The process of AI generalization is a process of driving enabling technology through systemic automation and intelligence. On the one hand, this is a result of the accumulation of human science and technology; on the other hand, it is an inevitability demanded by modern social production. AI generalization, on the one hand, makes the overall operation of the system, rather than the effective operation of parts, the unit of activity, thereby enhancing the integration of productive labor as a whole, making the era when automation "uses only one tool per worker" [3]3 This phrase echoes Marx's description in *Capital*, Vol. 1, of the cooperative workshop where many workers use a single tool, contrasted here with the integrated automation of AI. a thing of the past. On the other hand, it gives systems that are easy to replicate and migrate broader applicability, driving the organization of production and technological progress, weakening the status of original creators and users through iteration and optimization, and fundamentally resetting the organizational logic of labor's creation of value.

第一，人工智能通用化在升级取代技术的过程中，正在取代人类劳动的一般形式和过程，使劳动创造价值的形态与传统生产和自动化时代区别开来。人工智能的通用化过程，是以系统自动化与智能化驱动使能技术的过程，这一方面是人类科学与技术积累使然，另一方面则是现代社会生产要求的必然。人工智能的通用化，一方面使系统的整体运行，而非部分的有效运行成为活动的单位，从整体上提升了生产劳动的集成度，使自动化“机构只使用一个工具的人工”的时代一去不复返；另一方面，使易于复制与迁移的系统具有更广泛的适用性，推动生产的组织和技术进步，在迭代与优化中弱化原创者与使用者的地位，从根本上重置了劳动生产价值的组织逻辑。

Second, generalized AI is essentially a revival of overall labor capacity, a factual affirmation of the value of accumulated human technological labor, not a negation of labor's creation of value. We know that "AI is about making computers do the kinds of things that human intelligence can do." [4]4 This is a paraphrase of a common definition of AI, often attributed to John McCarthy, the coiner of the term "artificial intelligence." The generalization of AI means that AI "extends automation beyond the realm of routine work into some entirely new and unexpected areas." [5]5 This is a reference to the expanding scope of automation, as discussed in contemporary AI literature. Following this logic, AI turns the "enabling technology" originally driven by human routine labor into a new "replacement technology," to "perform the operations that workers used to perform with similar tools." [6]6 This phrasing adapts Marx's description of machinery replacing handicraft labor in *Capital*, Vol. 1. However, as a technological system, AI clearly cannot be separated from existing human material and spiritual conditions, nor can it be spontaneously created without human labor.

第二，通用化的人工智能本质上是对整体劳动能力的复活，是对人类技术积累劳动价值的事实肯定，而非对劳动创造价值的否定。我们知道，“人工智能就是使计算机做各种人类智能能做的事情”。人工智能的通用化，意味着人工智能“使得自动化超出了常规工作的范围，延伸到了一些全新的、意想不到的领域”。由此逻辑，人工智能使原本人类常规劳动驱动的“使能技术”成为全新的“取代技术”，来“完成过去工人用类似的工具所完成的那些操作”。然而，人工智能作为一种技术体系，显然无法离开人类既有的物质和精神条件，也不可能在没有人类劳动的前提下自发创建出来。

Third, although the generalization of AI, which "uses the same tools to expand human capabilities in mental, economic, social, and other aspects," [7]7 This is a paraphrase of the stated goals of general AI development, emphasizing augmentation rather than mere replacement. presents the directness of "replacing humans with increasingly powerful combinations of computer hardware and software," it raises the question of the value of the replacement technology created by humans. From the perspective of the comprehensiveness of AI generalization, AI seems to be fundamentally excluding the value of human activity. However, in the process of integrating enabling technology and replacement technology, AI generalization does not represent the emergence of a new labor subject, but rather the adoption of a new organizational form for human labor based on the accumulation of existing labor, generating new practical demands and forming a revolutionary production form.

第三，人工智能“使用相同的工具，在脑力、经济、社会等方面拓展人类能力”的通用化，虽然呈现出“用日益强大的计算机硬件和软件组合取代人类”的直接性，但是却提出了人类创造的取代技术的价值性问题。从人工智能通用化的全面性上讲，人工智能似乎正在从根本上排除人类活动的价值。但是，人工智能通用化在使能技术与取代技术一体化的过程中，并非全新的劳动主体的出现，而是人类劳动在既有劳动的积淀之中采用了全新的组织形态，产生了全新的现实需求，形成了革命性的生产形态。

AI's "Human-like Labor" and the Confirmation of Property Rights for Social Products

人工智能的“类人劳动”与社会产品的确权

Generalized AI, through human-like labor, "participates" in the cognition, transformation, and creation of objects, transferring the use value inherent in replacement technology to objectified social products. Although it is "unmanned" production in terms of production logic, the social products it outputs are still social products with increased use value. In its intelligent operation, general AI discovers, transforms, and innovates the use value of production objects. Based on the superposition of the use value of raw materials and AI, it achieves an increase in the quantity or even a qualitative change in use value. On the one hand, this is a concrete manifestation of AI's human-like labor capacity; on the other hand, it raises the difficult issue of confirming property rights for social products.

通用化的人工智能以类人劳动“参与”对象的认知、改造和创造，将蕴含于取代技术中的使用价值转移到对象化的社会产品之中，虽然在生产逻辑上是“无人化”的生产，但是产出的社会产品仍是使用价值增值的社会产品。通用人工智能在智能化的运行中发掘、改造和创新生产对象的使用价值，在原料与人工智能使用价值叠加的基础上实现了使用价值量的增加甚至质的改变，一方面是人工智能类人劳动能力的具体体现，另一方面更引发了对社会产品的确权难题。

First, the human-like labor of general AI is the product of system operation after the input of external purposes into the intelligent system, not the realization of the purpose of creating social products. Generalized AI not only elevates the logic of automation in explaining human physical strength, but also creatively achieves the liberation of mental and intellectual capacity. This does not mean the arrival of a singularity in the logic of social production, but rather a deconstruction of the meaning of labor's confirmation of property rights. This is because generalized AI itself is a virtualized, digitized, systematized, and intelligent "general condition of production." [8]8 "General condition of production" (一般生产条件) is a Marxist term for infrastructure and other broad preconditions for production, here applied to AI as a systemic condition. In its development toward generalization, AI not only achieves automation of the production process, intelligent production intervention, and closed-loop production regulation in the production of specific social products, enabling the output of use value driven solely by human instructions; it is also turning toward the automation and intelligence of "machine self-production," excluding the vast majority of workers from the production of the social machine through the closure of technological logic, the completeness of the intelligent system, and the effectiveness of intelligent iteration. However, generalized AI itself is merely a "revolution in productive forces manifested as a technological revolution," [9]9 This phrase is adapted from Marx's discussion of the relationship between technological revolution and productive forces in *Capital*. capable of effective output in response to input social needs, but it does not propose purposeful production.

第一，通用人工智能的类人劳动，是外部目的输入智能体系之后的系统运行产物，而非创造社会产品的目的性实现。通用化的人工智能不仅升华了自动化解释人类体力的逻辑，而且开创性地实现了对脑力和智力的解放。这并非意味着社会生产逻辑奇点的来临，而是对劳动确权意义的解构。因为，通用化的人工智能本身是虚拟化、数字化、系统化和智能化的“一般生产条件”。人工智能在通用化的发展中，不仅在具体社会产品生产中实现了生产过程的自动化、生产干预的智能化和生产调节的封闭化，依靠人类指令驱动就可以完成使用价值的产出；而且正在转向“机器自我生产”的自动化与智能化，其以技术逻辑的封闭性、智能系统的完备性和智能迭代的有效性，将绝大多数劳动者排除在社会机器的生产之外。但是，通用化的人工智能本身只是“表现为工艺革命的生产力革命”，其能够实现社会需要输入的有效产出，但没有提出目的性的生产。

Second, within the logic of generalization, AI is an integration and organization of existing human behavioral capacities, aimed at improving the efficiency of existing human capacities and avoiding the impact of human biological limitations on social production, rather than creating use value itself. It is undeniable that AI has brought about revolutionary changes to the social production logic of creating use value. It not only "effectively replaces labor power in the form of technology" [10]10 This is a paraphrase of Marx's analysis of machinery as a means of replacing labor power in *Capital*, Vol. 1. through automation, but also alleviates product shortages, limited production capacity, and insufficient growth through intelligence, and, through cross-domain generality, makes the production of use value more real-time, precise, and effective. Therefore, AI not only "defines a new world through technological revolution," [11]11 This is a common expression in discussions of the transformative impact of AI, often used in popular and academic contexts. but also defines a new model for the output of social products through general-purpose technology. In other words, AI's human-like labor makes the acquisition of use value more convenient, its types more abundant, and its forms more diverse, but it does not constitute the development of entirely new use values.

第二，人工智能在通用化的逻辑中是对人类既有行为能力的集成与有机化，其目的在于提高人类既有能力的效能，避免人类生物有限性等对社会生产的影响，而非创造使用价值本身。不可否认的是，人工智能使生产使用价值的社会生产逻辑发生革命性变革，既在自动化中“以技术的形式有效地替代劳动力”，又以智能化缓解产品的短缺、产能的有限和增长的不足，更以跨领域的通用性使得使用价值的生产更具有实时性、精准性和有效性。因此，人工智能不仅以“技术革命定义了一个新世界”，而且以通用技术定义社会产品产出的新模式。或者说，人工智能的类人劳动使使用价值的获取更为便利、种类愈加丰富、形态更加多元，但是并非全新使用价值的开发。

Third, the human-like labor of general artificial intelligence is the result of the “activation” of the capacity of the means of production; it is neither the subject that validates claims to social products, nor the sole basis for those who possess artificial intelligence to validate claims to social products. General artificial intelligence, in the form of “human-affiliated technology,” liberates or even replaces human labor, shifting the output of social products from the “factor-driven” model of the automation era to an “innovation-driven” one. This means that artificial intelligence cannot serve as the subject that validates claims, because only the laborers who achieve genuine innovation are the ontological foundation and the socioeconomic principle of innovative use-value. The generalized artificial intelligence “simulates all human capabilities—vision, reasoning, language, learning, etc.—to meet various challenges,” realizing human-like activity logic through technical logic. It is itself based on powerful computing power, massive data, advanced algorithms, and effective integration, originating from the animal-like instinctive labor capacity of humans. Therefore, it is a socio-historical achievement accumulated by historical laborers within the factors of production and the production environment, not the resurrection and subjectification of a material system.

第三，通用人工智能的类人劳动是生产工具“能力”活化的结果，既不是社会产品确权的主体，又不能作为占有人工智能者确权社会产品的唯一根据。通用人工智能以“属人技术”的形态解放人类劳动甚至代替人类劳动，使社会产品的产出由自动化时代的“要素驱动”转向“创新驱动”。这意味着，人工智能不能作为确权的主体，因为只有实现真正创新的劳动者才是创新性使用价值的存在论根据和社会经济原则。通用化的人工智能“模拟的所有人类能力——视觉、推理、语言、学习等——可应对各类挑战”，以技术逻辑的方式实现了类人的活动逻辑，本身是基于强大算力、海量数据、先进算法和有效集成，源于人的动物式的本能劳动能力。因此，它是既有历史劳动者积累于生产要素和生产环境中的社会历史性成果，而非物体系的复活和主体化。

The “Quasi-Subject” of Artificial Intelligence and the Enjoyment of Social Wealth

人工智能的“类主体”与社会财富的享有

Artificial intelligence, based on deep learning, generative principles, perceptual acquisition, and other methods, fully exploits and even constructs entirely new capabilities that surpass existing ones. In the “general automation of cognitive labor,” it “replaces the labor of ‘intellectuals,’” thereby acquiring autonomy in a logical sense and becoming the quasi-subject of the aforementioned human-like labor. The quasi-subject of artificial intelligence not only comprehensively changes the form of labor, enhances labor efficiency, and promotes the “withdrawal” of industrial workers from production, but also, through the “intelligent labor” of the quasi-subject, transforms the way social wealth is enjoyed.

人工智能基于深度学习、生成式原则、感知性获取等充分挖掘，甚至建构起超越既有能力的全新能力，在“认识劳动的普遍自动化”中“取代了‘知识分子’们的劳动”，从而在逻辑上具有了自主性，成为前述类人劳动的类主体。人工智能的类主体，不仅全面改变了劳动形式，提升了劳动效率，推进了产业工人的生产“离场”，而且更是以类主体的“智能劳动”改变着社会财富享有的方式。

First, the “intellectual value” of the artificial intelligence “quasi-subject” turns the means of labor into an autonomous industrial “perpetual motion machine.” Through the technical logic of “self-governance, self-reproduction, and self-cognition,” it extends a quasi-subject capacity “independent” of human behavior, reshapes actual material forces, and renews the mode of creating social wealth. The artificial intelligence “quasi-subject,” based on deep learning from big data, task-specific orientation, and the general integration of specialized artificial intelligence, becomes the optimal “human replacement” for artificial intelligence. On the one hand, it breaks the traditional restriction that production workers must engage in labor within the direct production process; on the other hand, it realizes certain unique and fundamental mechanisms of human thought, precisely grasps each link and field of wealth creation, and achieves personalized and visualized management of the labor process, thereby allocating factors of production more rationally.

第一，人工智能“类主体”的“智力价值”，使劳动资料成为自主性的工业“永动机”，其以“自我治理、自我繁衍、自我认知”的技术逻辑，延展出“独立”于人类行为的类主体能力，重塑了现实物质力量，更新了社会财富的创造方式。人工智能“类主体”基于大数据的深度学习、根据特定任务导向和专业人工智能的通用集成成为人工智能的最佳人体“人替”，一方面，打破了传统意义上生产工人必须在直接生产过程中从事劳动的人身限制；另一方面，实现人类思维某些独特而根本的机制，精确地掌握财富创造的各环节、各领域，实现劳动过程管理的个性化与可视化，从而更为合理地配置生产要素。

Second, the power beyond human capacity formed by the artificial intelligence “quasi-subject,” aside from the driving force of intelligent autonomous development, also continues the logic of fixed capital’s mechanization that excludes humans, remaining trapped within the logic of capital’s self-expansion. Reshaping the just logic of social wealth enjoyment is a prerequisite for resolving the current rift between the social production system and the value exchange system. “Intelligent autonomy,” becoming increasingly intelligent through evolution, causes artificial intelligence as a technology to erode the meaning of human labor. “Labor and work are no longer the ‘patent’ or unique essential activity of humanity.” The “technological unemployment” thus promoted, producing issues such as the “digital poor,” the “useless class,” and “universal proletarianization,” is clearly not an inevitable outcome of fetishism, but rather a mode of devouring the subject constructed by the new form of capitalist self-expansion.

第二，人工智能“类主体”形成的超越人的力量，除去智能化自主发展的推力之外，也延续了固定资本机器化排斥人的逻辑，依然困囿于资本增殖的逻辑之中。重塑社会财富享有的正义逻辑是解决当下社会生产系统与价值交换系统之间撕裂的前提。“智能化的自主”在进化中越来越聪明，使人工智能作为技术取代正在侵蚀人类劳动的意义。“劳动、工作已经不再是人类的‘专利’、特有的本质性活动”，由此推进的“技术性失业”产生的诸如“数字穷人”“无用阶级”“普遍的无产阶级化”问题，显然不是物神论的必然，而是资本主义增殖新形态所建构的吞噬主体的方式。

Third, the artificial intelligence “quasi-subject” raises the question of human essence in an ontological sense. As an objectified product of humanity’s essential power, realizing the intelligent return of artificial intelligence to the human essence, and making it a socialist “common wealth,” is the only way to ensure the economic justice and political rationality of artificial intelligence development, and to guarantee that it serves human free development. As a comprehensive manifestation of the upgrade of human capacity and intelligence, the generalization of artificial intelligence precisely embodies the qualitative enhancement of humanity’s “subjective creativity” over artificial intelligence—that is, the full realization of human capacity to “apply intelligence,” “master intelligence,” and “realize intelligence.”

第三，人工智能“类主体”在存在论意义上提出了人的本质问题，作为人的本质力量的对象化产物，实现人工智能对人的本质的智能化复归，使其成为社会主义“共同的财富”才能确保人工智能发展的经济正义性、政治合理性，并确保其服务于人的自由发展。人工智能作为人类能力、智力升级的全面体现，它的通用化恰恰体现了人类对于人工智能的“主体创制性”得到质的提升，也即代表人类“应用智能”“掌握智能”“实现智能”的人的能力全面实现。

(The author is Vice Dean and Professor of the School of Marxism at South China Normal University, and a Distinguished Researcher of the Guangdong Provincial Research Center for

（作者系华南师范大学马克思主义学院副院长、教授，广东省习近平新时代中国特色社会主义思想研究中心特约研究员）

Online editor: Huihui

网络编辑：慧慧