Artificial Intelligence and the Transformation of the Capitalist Mode of Production

The capitalist mode of production has always been a main subject of Marxist political economy. In the preface to the first edition of *Capital*, Volume 1, Marx clearly states: “What I have to examine in this work is the capitalist mode of production, and the relations of production and exchange corresponding to it.” In *Capital*, Volume 2, Marx further states: “Whatever the social form of production, labourers and means of production always remain its factors. But while they are in a state of separation from each other, both of them are factors in production only in possibility. For production to go on at all, they must be combined. The special manner in which this combination is accomplished distinguishes the different economic epochs of the social structure.” Marx’s analysis in *Capital*, especially Volume 1, is precisely carried out with the capitalist mode of production as its object, including the production of absolute surplus value and relative surplus value, the process of capital accumulation, and so on. Marx scientifically proved that the basis of the capitalist mode of production is wage labour, and that under this mode of production, “all methods for raising the social productivity of labour are brought about at the cost of the individual worker; all means for the development of production transform themselves into means of domination over, and exploitation of, the producers.”

资本主义生产方式一直是马克思主义政治经济学探讨的主要内容。马克思在《资本论》第1卷第一版序言中明确指出：“我要在本书研究的，是资本主义生产方式以及和它相适应的生产关系和交换关系。”在《资本论》第2卷中，马克思又讲道：“不论生产的社会的形式如何，劳动者和生产资料始终是生产的因素。但是，二者在彼此分离的情况下只在可能性上是生产因素。凡要进行生产，它们就必须结合起来。实行这种结合的特殊方式和方法，使社会结构区分为各个不同的经济时期。”马克思的《资本论》特别是《资本论》第1卷的分析正是以资本主义生产方式为对象展开的，包括绝对剩余价值的生产和相对剩余价值的生产、资本的积累过程等。马克思科学地证明了，资本主义生产方式的基础是雇佣劳动，在这种生产方式下，“一切提高社会劳动生产力的方法都是靠牺牲工人个人来实现的；一切发展生产的手段都转变为统治和剥削生产者的手段”。

With the development of the productive forces, the capitalist mode of production is constantly transformed; both the labour force and the means of labour, as well as the mode of capital’s employment of wage labour that results from their combination, are in a state of continuous change. Historically, the first scientific and technological revolution, driven by the steam engine, brought human society into the steam age, laying a powerful material foundation for the capitalist mode of production and establishing the dominant position of capital over labour. The second scientific and technological revolution, driven by the electric motor, brought human society into the electrical age, accelerating the concentration and centralisation of capital and strengthening capital’s control over and exploitation of labour. The third scientific and technological revolution, driven by the computer, brought human society into the information age, significantly increasing the pace of transformation of the means of labour. Information technology enabled automation in production, and the role of scientific and technological means in production increasingly surpassed that of living labour. Entering the 21st century, the emergence of artificial intelligence has driven the arrival of a new round of scientific and technological revolution. Artificial intelligence is more powerful than any driving force since the Industrial Revolution and is profoundly changing the capitalist mode of production. In the age of artificial intelligence, where capitalism is heading depends to a large extent on where the transformation of the capitalist mode of production is heading. In view of this, foreign academic circles have engaged in extensive discussions on issues related to artificial intelligence and the transformation of the capitalist mode of production in recent years. This article aims to review these discussions and analyse them using the perspective of Marxist political economy.

随着生产力的发展，资本主义生产方式不断变迁，无论是劳动力、劳动资料还是二者相结合而产生的资本雇佣劳动的方式都处于持续的变化之中。在历史上，由蒸汽机推动的第一次科技革命使人类社会进入蒸汽时代，为资本主义生产方式奠定了强大的物质基础，确立了资本对劳动的统治地位。由电机推动的第二次科技革命使人类社会进入电气时代，加速了资本的积聚与集中，强化了资本对劳动的控制与剥削。由计算机推动的第三次科技革命使人类社会进入信息时代，劳动资料的变革速度显著提升。信息技术使生产实现了自动化，科技手段在生产中的作用日益高于活劳动的作用。进入21世纪，人工智能的出现推动了新一轮科技革命的到来。人工智能比工业革命以来的任何推动力都要强大，正在深刻地改变着资本主义生产方式。在人工智能时代，资本主义向何处去在很大程度上取决于资本主义生产方式的变迁向何处去。有鉴于此，国外学界围绕近年来人工智能与资本主义生产方式变迁的相关问题展开了大量讨论。本文旨在对这些讨论进行梳理，并运用马克思主义政治经济学的观点进行分析。

I. The Transformation of the Machine System by Artificial Intelligence

一、人工智能对机器体系的改造

The distinction between economic epochs lies not in what is produced, but in how it is produced, and by what means of labour it is produced. “The revolution in the mode of production, in manufacture, takes as its starting point the labour power; in large-scale industry, it takes as its starting point the means of labour.” In *Capital*, Marx analyses in detail the history of the development of the machine, the main carrier of the means of labour, pointing out that “all fully developed machinery consists of three essentially different parts: the motor mechanism, the transmitting mechanism, and the tool or working machine. The motor mechanism is the driving force of the whole mechanism.” The transmitting mechanism “regulates the motion, changes its form where necessary (for example, from linear to circular), and distributes and transmits it to the tool machines. The function of these two parts of the machinery is merely to impart motion to the tool machine, which then seizes upon the object of labour and modifies it according to a definite purpose.” Marx’s analysis of the machine reveals the essential characteristics of the development of the means of labour since the Industrial Revolution. On this basis, the third industrial revolution, centred on computer technology in the 1960s, pushed the machine system to a new stage of automation. The distinctive feature of the automatic machine is the addition of a new device—the automatic control system (the control machine)—to the basic composition of the tool machine, the motor mechanism, and the transmitting mechanism. “The third industrial revolution, starting from the control machine, made complex processes controllable, using machines to replace part of the functions of the human brain.” The control machine can process received information and adjust the machine’s motion according to changes in the external environment. Ramin Ramtin points out that the automatic machine is a production system integrating control technology, composed of a specific production structure (i.e., traditional mechanised technology) and an information utilisation module. An important feature of the automatic machine is the electronic synchronisation of the processing procedures of the mechanised machine. Automation is a new revolution in the development process of the means of labour; it opens a new stage in which machines replace the human brain in processing information, greatly promoting the transformation of the mode of production. However, this is only the initial stage of the digital revolution and still has obvious limitations. In fact, automatic machines only replace the labour of workers in certain links; they do not replace all operational workers in the production process. The operation of automatic machines is inseparable from the care and correction of workers, and even more so from the instructions issued by enterprise management; thus, problems such as rigidity, inflexibility, and insufficient precision exist.

各经济时代的区别不在于生产什么，而在于怎样生产，用什么劳动资料进行生产。“生产方式的变革，在工场手工业中以劳动力为起点，在大工业中以劳动资料为起点。”在《资本论》中，马克思详细地分析了劳动资料的主要载体——机器的发展历史，指出“所有发达的机器都由三个本质上不同的部分组成：发动机，传动机构，工具机或工作机。发动机是整个机构的动力”，传动机构“调节运动，在必要时改变运动的形式（例如把垂直运动变为圆形运动），把运动分配并传送到工具机上。机器的这两个部分的作用，仅仅是把运动传给工具机，由此工具机才抓住劳动对象，并按照一定的目的来改变它”。马克思关于机器的分析揭示了工业革命以来劳动资料发展的本质特征。在此基础上，1960年代发生的以计算机技术为核心的第三次工业革命将机器体系推向了自动化新阶段。自动化机器的显著特点是在工具机、动力机和传动机的基础构成上增添了一个新的装置——自动控制系统（控制机）。“第三次工业革命，从控制机开始，让复杂的过程变成可控的，利用机器代替人脑的部分功能。”控制机能够对接收到的信息进行处理，根据外部环境变化调节机器的运动。拉明·拉姆廷（Ramin Ramtin）指出，自动化机器是一个集成了控制技术的生产系统，由特定的生产结构（即传统的机械化技术）和信息利用模块所组成。自动化机器的重要特征是将机械化机器的加工工序进行电子化同步。自动化是劳动资料发展进程中的一次新革命，它开启了机器代替人脑处理信息的新阶段，大大推动了生产方式的变革。但是，这只是数字化革命的初级阶段，还存在明显的局限。事实上，自动化机器只是取代某些环节的工人劳动，并不会取代生产过程中的所有操作工人。自动化机器的运转离不开工人的照料和纠偏，更离不开企业管理层下达的指令，因而存在僵化、不灵活、精细化程度不高等问题。

The emergence of artificial intelligence pushes the digital revolution to the next stage, and also pushes the progress of the means of labour to a new stage—the stage of intelligentisation. The intelligent transformation of machines by artificial intelligence overcomes the drawbacks of automatic machines. In a sense, the essence of artificial intelligence is the intelligentisation of the control machine, endowing traditional machines with capabilities they originally lacked, such as data mining, data analysis, automatic programming, intelligent control, machine vision, and deep learning. As a “computational agent capable of generating intelligent behaviour,” artificial intelligence can perform a series of comprehensive analyses similar to the thinking process of the human brain, based on its environment and circumstances. Therefore, artificial intelligence equipment equipped with sensors and analysers can monitor and correct working conditions in real time, possessing, in a sense, the learning and perceptual abilities unique to humans. Naoyuki Nomura believes that artificial intelligence achieves comprehensive optimisation and surpassing of automatic machines, enabling rapid extraction, processing, and analysis of large amounts of information data.

人工智能的出现将数字化革命推向了下一个阶段，也将劳动资料的进步推向了新的阶段，即智能化阶段。人工智能对机器的智能化改造克服了自动化机器的弊病。从某种意义上讲，人工智能的本质就是控制机智能化，使传统机器拥有其原先所不具备的数据挖掘、数据分析、自动程序设计、智能控制、机器视觉、深度学习等能力。人工智能作为“能够产生智能行为的计算代理人”，可以根据其所处环境与情形进行一系列类似于人类大脑思考过程的综合分析。因此，配有传感器和分析器的人工智能设备可以实时监测和修正工作状态，在某种意义上拥有人类特有的学习能力和感知能力。野村直之认为，人工智能对自动化机器实现了全面的优化和超越，能够对大量的信息数据进行快速提取、处理和分析。

So, what impact does the application of intelligent machines have on the capitalist mode of production? Relevant research in foreign academic circles indicates that, due to the transformation by artificial intelligence, the capitalist mode of production has undergone changes in two aspects.

那么，智能化机器的运用对资本主义生产方式产生了哪些影响呢？国外学界的相关研究表明，由于人工智能的改造，资本主义生产方式发生了两方面的变化。

On the one hand, the platform has become an important form of production organization. Nick Srnicek points out that after the 2008 financial crisis, artificial intelligence was widely applied, significantly reducing the technical cost of converting people’s simple activities into data, and data became the core resource of social production. Under these circumstances, the capitalist mode of production organization developed a new form—the platform. On the platform, different actors from different regions, such as customers, advertisers, producers, suppliers, and laborers, can interact with information or conduct economic activities at any time and any place. The characteristics of the platform determine that it has strong network effects and a tendency toward monopoly. As for network effects, the larger the platform’s scale and the more economic activities it hosts, the greater its value, and the greater the utility platform users can obtain. The tendency toward monopoly means that platform capitalists have a strong desire to expand through mergers, acquisitions, and other means, in order to secure a larger market, a higher market position, and more autonomous pricing power. Thus, both platform capitalists and platform users are willing to see the rapid development and expansion of the platform. Scholars such as Jamie Woodcock argue that the platform has become the center of people’s production and social activities; it can bring users together, better match the supply and demand of various commodities, and regulate nearly all conceivable economic activities. Specifically, the platform can provide ample flexibility and convenience for various types of users: for example, employers can freely choose the time and method of hiring workers, customer needs can be quickly realized and satisfied, and workers are not forced by traditional employment contracts but can freely choose their working hours, work locations, and work content.

一方面是平台成为生产组织的重要形式。尼克·斯尔尼塞克（Nick Srnicek）指出，在2008年金融危机之后，人工智能获得了广泛应用，它使人们的简单活动转化为数据的技术成本显著降低，数据成为了社会生产的核心资源。在这种情况下，资本主义的生产组织形式发展出了一种新的形式——平台。在平台上，客户、广告商、生产商、供应商、劳动者等来自不同地域的不同行为主体能够在任何时间、任何地点进行信息交互或开展经济活动。平台的特点决定了其具有很强的网络效应和垄断倾向。就网络效应而言，平台的规模越大、承载的经济活动越多，它的价值就越大，平台用户能够得到的效用也就越大。而垄断倾向是指，平台资本家拥有强烈的意愿通过兼并、收购等途径进行扩张，以便使自己拥有更大的市场、更高的市场地位以及更自主的定价权力。从而，无论是平台资本家还是平台用户，都愿意看到平台的快速发展与扩张。杰米·伍德科克（Jamie Woodcock）等学者认为，平台已经成为人们进行生产活动和社交活动的中心，它能够将用户聚集在一起，使各种商品的供应与需求得到更好的匹配，可以调节近乎所有可以想象到的经济活动。具体而言，平台能够为各类用户提供充足的灵活性和便利性，譬如雇主可以自由选择雇佣工人的时间、方式，客户需求能够被快速地实现与满足，工人不必被传统的雇佣合同所逼迫，而是可以自由地选择工作时间、工作地点和工作内容。

On the other hand, the “gig economy” model has emerged. The emergence of the platform has brought about significant changes in the mode of employment. In the past, the combination of labor and capital occurred within the enterprise through centralized management. Now, through the platform, workers can relatively freely and independently sign temporary work agreements with employers outside the enterprise. Thus, an informal, unstable, temporary, and fragmented mode of labor organization based on online platforms has gradually taken shape, namely the “gig economy.” Woodcock argues that the gig economy specifically includes two different types of work models. The first is the gig-ification of traditional work. Artificial intelligence can track, analyze, and predict workers’ work status and daily behavior in real time; what it changes is not the traditional work content but the organizational methods and processes. This is because occupations such as construction workers, repairmen, food delivery riders, taxi drivers, and domestic cleaners already existed widely before the emergence of platforms. The second is the new types of work generated alongside artificial intelligence, so-called “ghost work.” Why is it called “ghost work”? Mary L. Gray and Siddharth Suri point out that, in order to continuously improve artificial intelligence, capitalists instruct software engineers to publish a large number of “micro-tasks” through the platform, such as training models and information annotation. Workers receive instructions and submit work results through the Application Programming Interface (API), and receive a certain amount of remuneration automatically settled by artificial intelligence. The entire process is as invisible and intangible as a ghost, hence the name “ghost work.” Through the “ghost work” model, the new assembly line generated by artificial intelligence brings together workers scattered across the globe, forming a collective labor force. What the new assembly line produces is no longer individual parts, but individual “task sets.” The division of labor in the traditional assembly line is replaced by a large number of crowdsourcing and outsourcing arrangements.

另一方面是“零工经济”模式的出现。平台的出现使雇佣方式发生了重要改变。过去，工人与资本的结合是在企业之内，通过集中化的管理完成的。而现在，通过平台，工人在企业之外就可以相对自由且独立地与雇主签订临时工作协议。由此，一种基于在线平台的非正式、不稳定、临时性、零散的劳动力组织模式逐渐形成，即“零工经济”（gig economy）。伍德科克认为，零工经济具体包含两类不同的工作模式，其一是零工化的传统工作。人工智能可以实时跟踪、分析并预测工人的工作状态和日常行为，其改变的不是传统的工作内容而是组织方式和流程。因为建筑工、维修工、外卖送餐员、出租车司机、家政保洁员等职业在平台出现前就已经广泛存在。其二是伴随人工智能而产生的新的工作种类，即所谓的“幽灵工作”（ghost work）。为什么称为“幽灵”呢？玛丽·L.格雷（Mary L.Gray）和西达尔特·苏里（Siddharth Suri）指出，为了不断完善人工智能，资本家会命令软件工程师通过平台发布大量的“微任务”，譬如训练模型、信息标注等。工人通过应用程序接口（Application Programming Interface, API）接受指令、反馈工作成果，并获得由人工智能自动结算的一定数额的报酬。整个过程如同幽灵一样看不见、摸不到，“幽灵工作”便由此得名。通过“幽灵工作”模式，人工智能生成的新流水线把分散在全球各地的工人汇聚起来，形成了集体劳动力。新流水线生产的不再是一个个零件，而是一个个“任务集”。传统流水线上的分工形式被大量的众包和外包所替代。

II. The Impact of Artificial Intelligence on Laborers

二、人工智能对劳动者的影响

Laborers are one of the key elements of the productive forces. Under the capitalist mode of production, laborers are in a position dominated by capital. Relevant analysis shows that the application of artificial intelligence has not changed, but rather strengthened, this dominated position, mainly reflected in three aspects: first, the application of artificial intelligence has given rise to a new form of labor—digital labor; second, the application of artificial intelligence has promoted the polarization of the labor force; third, the degree of unity among the working class has been weakened.

劳动者是生产力的关键要素之一。在资本主义生产方式下，劳动者处于被资本支配的地位。相关分析表明，人工智能的应用不仅没有改变、反而强化了这种被支配的地位，主要体现在三个方面：一是人工智能的应用催生了新的劳动形态——数字劳动；二是人工智能的应用促进了劳动力的两极分化；三是工人阶级的团结程度被削弱。

First, artificial intelligence has given rise to a new form of labor, namely digital labor. In digital labor, capital can appropriate almost all of the laborer’s work without compensation. Christian Fuchs argues that, driven by intelligent technology, a new form of labor different from traditional industrial labor but equally serving capital accumulation has emerged—digital labor. The most typical digital laborers are users of social media platforms such as Facebook, Google, and Twitter. The daily browsing and clicking behaviors of social media users possess both consumption and production attributes. In terms of consumption attributes, social media users’ behaviors such as browsing, watching, and purchasing can bring them entertainment and relaxation experiences. In terms of production attributes, the daily behaviors of social media users are fully recorded by algorithms, producing two important commodities for the operators of social media—data and attention. In the view of Fuchs and others, the operators of social media do not need to pay wages to digital laborers, yet they can comprehensively collect and appropriate the products of digital laborers without compensation, and earn huge profits by selling them or placing advertisements. Hector Postigo, using the video website YouTube as an example, discusses the labor form of digital laborers on social media. He points out that social media platforms like YouTube have constructed a complex virtual architecture, including comment systems, tagging systems, advertising systems, video ranking systems, game control systems, and image and video capture systems. This architecture possesses social, entertainment, and incentive functions, causing a large number of digital laborers to become immersed in it, investing a great deal of personal leisure time and money, tirelessly watching advertisements, creating content, and generating data.

第一，人工智能催生了新的劳动形态，即数字劳动。在数字劳动中，资本可以近乎无偿地占有劳动者的全部劳动。克里斯蒂安·富克斯（Christian Fuchs）认为，在智能化技术的推动下，出现了一种不同于传统产业劳动、但同样服务于资本积累的新的劳动形态——数字劳动。最典型的数字劳动者就是脸书、谷歌、推特等社交媒体上的用户。社交媒体用户的日常浏览和点击行为兼具了消费和生产两种属性。就消费属性而言，社交媒体用户的浏览、观看、购买等行为能够为他们自身带来消遣和放松的体验；而就生产属性而言，社交媒体用户的日常行为会被算法全盘记录，并为社交媒体的运营商生产两种重要商品——数据和注意力。在富克斯等人看来，社交媒体的运营商无需向数字劳动者支付劳动报酬，却能全面收集并无偿占有数字劳动者的产品，并以售卖或投放广告的形式赚取巨额利润。赫克托·波斯蒂戈（Hector Postigo）以视频网站优兔为例，讨论了社交媒体上的数字劳动者的劳动形态。他指出，优兔等社交媒体搭建起了一套复杂的虚拟架构，包括评论系统、标签系统、广告系统、视频排名系统、游戏控制系统和图像视频捕获系统等。这套架构兼具社交性、娱乐性和激励性，使大量的数字劳动者沉醉其中，投入大量的个人闲暇时间和金钱，不知疲倦地观看广告、创作内容并生成数据。

Second, the application of artificial intelligence has led to a polarization of the labor force, exacerbating social inequality. If workers are divided into three levels—high, middle, and low—based on knowledge reserves and skill levels, the application of AI has achieved large-scale substitution of middle-level workers and has led to skill polarization and income polarization between high- and low-level workers, intensifying social inequality. Research by David Autor and others shows that routine jobs are most severely impacted by AI, while high-skilled cognitive jobs and low-skilled manual jobs are relatively less affected. The most heavily impacted are middle-level workers such as warehouse managers, couriers, financial cashiers, ordinary construction workers, and translators. The skills these workers possess are highly substitutable, and they may be forced to lose their jobs due to the application of AI. Second, AI has caused severe polarization between high- and low-level workers and has placed new demands on their own capabilities. For mental workers with higher education, rich knowledge reserves, and professional technical skills, they can maintain relatively high labor remuneration and higher social status, but they are also required to master comprehensive algorithmic programming skills, data analysis capabilities, overall management and coordination abilities. Low-skilled manual workers engaged in simple physical and service work, on the other hand, are confined to corresponding workplaces, performing monotonous tasks for long periods and earning low wages—such as cleaners, gardeners, restaurant waiters, and security guards. Low-skilled work also includes data mining, data processing, and information labeling tasks that have emerged due to the widespread application of AI. For cost and corporate reputation considerations, capitalists typically outsource these low-skilled jobs to workers in underdeveloped countries or regions. The hourly wages of these outsourced workers are very low, ranging from a few dollars to a few cents per hour.

第二，人工智能的应用导致了劳动力的两极分化，加剧了社会不平等。如果根据知识储备和技能水平将劳动者分为高、中、低三个层次，人工智能的应用实现了对中层劳动者的大量替代，并导致了高、低两个层次劳动者的技能极化和收入极化，加剧了社会不平等。大卫·奥特（David Autor）等人的研究表明，常规性质的工作受人工智能的冲击最大，高技能的认知型工作和低技能的体力型工作所受的影响相对较小。受冲击最大的是仓库管理员、快递员、财务出纳员、普通建筑工人、翻译员等中层劳动者。这类劳动者掌握的技能的可替代性强，会因人工智能的应用而被迫下岗失业。其次，人工智能使高、低两个层次的劳动者发生了严重的两极分化并对他们自身的能力提出了新的要求。对于拥有较高学历、丰富的知识储备和专业技术能力的脑力劳动者而言，他们能够维持相对较高的劳动报酬、更高的社会地位，同时也被要求掌握完备的算法编程能力、数据分析技能、统筹管理和协调能力。而从事简单的体力工作和服务性工作的低技能体力型劳动者则被圈定在相应的工作场所，长期从事枯燥的工作、赚取较低的工资，譬如清洁工、园丁、餐厅服务员、保安等。低技能工作还包括由于人工智能的广泛应用而出现的数据挖掘、数据处理、信息标注等工作。出于成本和企业声誉的考量，资本家通常会将这些低技能工作外包给不发达国家或地区的劳动者去完成。这些外包劳动者的时薪非常低，大约从每小时几美元到几美分不等。

Third, the solidarity of the working class has been weakened. Behind the superficial prosperity of intelligent technology lies the increasingly isolated situation of large numbers of workers. The new forms of production organization brought by AI make it difficult for workers in many industries to meet and communicate with other workers during the work process, leading to looser connections among workers. Gray and others point out that “ghost work” has replaced traditional work models, work hours have become more flexible, and workers must compete with others to obtain job opportunities. Workers no longer share workplaces, work hours, or occupational identities. Due to the lack of common interests and solidarity mechanisms, workers cannot defend their labor prices. In a “race to the bottom,” workers who offer lower bids are often able to secure jobs. Scholars such as Iman Kumar Mitra argue that the labor market has seen both formal and informal forms of division, and capitalists have changed their customary management methods, further strengthening control and division over workers. Due to the application of AI, capitalists have given workers the illusion that they can earn more returns through individual effort. The influence of individualism in society has gradually become prominent. Capitalists rely on new forms of production organization to assign large amounts of work to independent workers, such as ride-hailing drivers, shared-housing hosts, and food delivery riders. Independent workers are not subject to minimum wage or employment tax constraints; as long as they invest more personal time in work, they can hope to earn more pay. This weakens connections among workers, severely impacts union work, and continuously reduces their power to confront capitalists. Valeria Pulignano also points out that fragmented work may bring instability to social structures. Work originally allowed workers to unite organically, but the fragmented “task sets” driven by AI have dissolved the tightly united proletariat.

第三，工人阶级的团结程度被削弱。在智能化技术表面繁荣的背后是大量劳动者日益孤立的境遇。人工智能带来的新生产组织形式使很多行业的劳动者在工作过程中难以同其他劳动者见面和交流，这就导致劳动者之间的联结变得更加松散。格雷等人指出，“幽灵工作”替代了传统的工作模式，工作时间变得更富弹性化，劳动者为了获得工作机会还需要同其他人竞争。劳动者不再共享工作地点、工作时间、职业身份。由于缺乏共同利益和团结机制，劳动者无法维护自己的劳动力价格。在“逐底竞争”中，出价较低的劳动者往往能够获得工作机会。伊曼·库马尔·米特拉（Iman Kumar Mitra）等学者认为，劳动力市场出现了正式和非正式两种形式的分化，资本家改变了以往惯用的管理手段，进一步加强了对劳动者的控制与分化。由于人工智能的应用，资本家给予了劳动者一种错觉，即让他们认为通过个人努力就可以赚取更多的回报。个人主义在社会中的影响逐渐凸显。资本家依托新的生产组织形式将大量的工作交由独立劳动者去执行，如网约车司机、共享民宿房主、外卖配送员等。独立劳动者不受最低工资、就业税等约束，只要将更多的个人时间投入工作就有希望赚得更多报酬。这使劳动者之间的联结变得弱化，工会工作受到巨大冲击，其对抗资本家的力量被不断削减。瓦莱里亚·普利尼雅诺（Valeria Pulignano）也指出，碎片化工作可能会给社会结构带来不稳定因素。工作本可以让工人有机地团结在一起，但是人工智能催生的碎片化“任务集”将紧密团结的工人无产阶级瓦解了。

III. New Characteristics of Capitalist Labor-Capital Relations

三、资本主义劳资关系的新特征

Engels pointed out, “The relation of capital and labor is the axis around which our entire modern social system revolves,” and it is also the core of the capitalist mode of production. Regarding the main characteristics of the relationship between capital and labor under the capitalist mode of production, Marx once summarized: “The worker works under the control of the capitalist, and his labor belongs to the capitalist”; “The product is the property of the capitalist, not the property of the direct producer, the worker.” AI has not changed the essential characteristics of the capitalist mode of production; instead, it has used new technologies and means to strengthen capital’s control and exploitation of labor. Relevant research in foreign academic circles discusses three main aspects of the new characteristics of capitalist labor-capital relations under intelligent conditions: first, capital has gained more and more advanced means to strengthen control over labor, known as “surveillance capitalism”; second, traditional workers have transformed into “ghost workers,” and the degree of exploitation suffered by workers has further increased; third, the contradiction between labor and capital has generally become hidden.

恩格斯指出，“资本和劳动的关系，是我们全部现代社会体系所围绕旋转的轴心”，也是资本主义生产方式的核心。对于资本主义生产方式下资本与劳动的关系的主要特征，马克思曾做过这样的概括：“工人在资本家的监督下劳动，他的劳动属于资本家”，“产品是资本家的所有物，而不是直接生产者工人的所有物”。人工智能并没有改变资本主义生产方式的本质特征，而是用新的技术和手段强化了资本对劳动的控制与剥削。国外学界的相关研究讨论了智能化条件下资本主义劳资关系的新特征，主要有三个方面：一是资本获得了更多、更先进的手段来加强对劳动的控制，即所谓的“监视资本主义”（surveillance capitalism）；二是传统的工人转变为“幽灵工人”，劳动者遭受的剥削程度进一步提升；三是劳动与资本之间的矛盾总体走向隐匿化。

(I) Capital’s Control over Labor Has Significantly Strengthened

（一）资本对劳动的控制显著增强

The emergence of large-scale machine industry greatly strengthened capital’s control over labor, “making the worker crippled, turning him into a partial man, degrading him to an appendage of the machine, subjecting him to the torment of labor, thereby emptying labor of content, and, as science is incorporated into the labor process as an independent force, alienating the intelligence of the labor process from the worker.” On the surface, with the widespread application of AI in various fields, capital’s control over labor seems to have loosened—for example, more flexible working hours, freer workplaces, workers being able to work alone or form temporary teams with other workers for collaboration, and even holding multiple paid jobs part-time. But in reality, AI has further strengthened capital’s control over labor, and the degree to which living labor is dominated by objectified labor has deepened. Shoshana Zuboff uses “surveillance capitalism” to characterize this feature. She explains that it is called “surveillance capitalism” because, in order to secure monopoly and dominance over data, capitalists use intelligent technological means to seek, mine, and capture information that originally belonged to the private sphere, and the entire process prominently exhibits the characteristics of surveillance. Zuboff refers to capitalists engaged in the economic activities of “surveillance capitalism” as “surveillance capitalists” and argues that all actions of surveillance capitalists consistently revolve around two main purposes: first, to strengthen comprehensive monitoring of workers during the labor process, enhance control over labor, and expropriate, steal, redistribute, or hoard surplus labor; second, to replace workers with AI, continuously expand the size of the industrial reserve army in society, and create an ever-larger “reservoir” of labor. Surveillance capitalists in a dominant position possess substantial capital and can occupy and consolidate absolute dominance in areas such as rule-making and labor hiring. Specifically, the strengthening of capital’s control over labor is reflected in the following three aspects.

机器大工业的出现极大地加强了资本对劳动的控制，“使工人畸形发展，成为局部的人，把工人贬低为机器的附属品，使工人受劳动的折磨，从而使劳动失去内容，并且随着科学作为独立的力量被并入劳动过程而使劳动过程的智力与工人相异化”。从表面上看，随着人工智能在各领域的广泛应用，资本对劳动的控制似乎松动了，比如劳动时间更加灵活、工作地点更加自由，劳动者可以单独工作或者与其他劳动者组成临时团队开展协作，甚至可以兼职从事多份有报酬的工作。但实际上，人工智能使资本对劳动的控制进一步增强了，活劳动受对象化劳动支配的程度进一步加深。肖莎娜·祖博夫（Shoshana Zuboff）用“监视资本主义”来刻画这一特征。她解释道，之所以叫“监视资本主义”，是因为资本家为争取到数据的垄断权和主导权，借助于智能化技术手段去寻找、挖掘和捕捉原本属于私人领域的信息，整个过程突出地呈现了监视的特点。祖博夫将从事“监视资本主义”经济活动的资本家称作“监视资本家”（surveillance capitalist），并认为监视资本家的所有行为始终围绕着两个主要目的进行：一是强化对劳动过程中劳动者的全方位监督，增强对劳动的控制，并对剩余劳动进行征用、盗取、重新分配或囤积；二是用人工智能替代劳动者，不断扩充社会中的产业后备军数量，创造一个越来越大的劳动力“蓄水池”。处于优势地位的监视资本家拥有雄厚的资本，可以在规则制定和劳动力雇佣等方面居于并巩固绝对主导地位。具体而言，资本对劳动的控制增强体现在以下三个方面。

First, capitalists have acquired new means of remote and comprehensive control. Bernard Stiegler argues that artificial intelligence has brought capitalism into the “age of hyper-control.” In his view, the integration of AI with internet technology has led to large-scale “de-localization” of production sectors, enabling capitalists to construct a vast global factory capable of remote control, making it common to employ workers worldwide for remote work. Capitalists who master AI can capture traces of individual and collective behavior from workers and consumers, collecting, analyzing, and storing relevant data, thereby strengthening control over labor and manipulation of consumer markets. From the perspective of workers, the workplace is no longer confined to the factory; it can be private residences, cafes, restaurants, or even vehicles. Scholars such as Rosalind Gill refer to these non-traditional workplaces as the “social factory.” In their view, the “social factory” includes both traditional factories and all locations where workers can perform labor. Within the “social factory,” not only working hours but the entire life course of the worker serves the process of capital accumulation. The “social factory” originates in the sphere of production but is not limited to it; it gradually spreads and profoundly transforms all spheres, including distribution, exchange, and consumption.

第一，资本家获得了远程与全面控制的新手段。贝尔纳·斯蒂格勒（Bernard Stiegler）认为，人工智能将资本主义带入了“超级控制时代”（the age of hyper-control）。在他看来，人工智能与互联网技术的融合使生产部门大规模地“去本地化”（de-localize），资本家构建起了一个能够实现远程控制的庞大世界工厂，使雇佣全球各地的劳动者进行远程工作成为一种普遍现象。掌握人工智能的资本家能够通过捕获劳动者和消费者的个人和集体行为痕迹，对相关数据加以收集、分析和存储，从而加强对劳动的控制以及对消费市场的操控。从劳动者的角度而言，劳动场所不再局限于工厂，可以是私人住宅、咖啡厅、餐馆甚至是交通工具。罗莎琳德·吉尔（Rosalind Gill）等学者将这些非传统意义上的工作场所称为“社会工厂”（social factory）。在他们看来，“社会工厂”是包括传统工厂以及劳动者能够进行工作的所有地点。在“社会工厂”中，不仅工作时间，劳动者全部的生命历程都将服务于资本积累过程。“社会工厂”产生于生产领域，但不限于生产领域，它会逐渐蔓延并深刻改造分配、交换、消费等各领域。

Second, capital monopolizes data factors. On the surface, unlike traditional factors of production such as capital, labor, and land, data factors as an important factor of production in the digital age exhibit characteristics such as incomplete exclusivity and incomplete competition. In reality, however, this is not the case. The sources of data factors are extremely abundant. When people use online platforms and social media, they unknowingly contribute personal digital labor to capitalists, producing vast amounts of data. Capitalists can appropriate the labor products of digital workers almost gratuitously and expropriate surplus value. Zuboff points out that whether it is people’s browsing records online, the content they post, or their personal biometric data, all can be secretly captured and privately appropriated by surveillance capitalists. Capitalists who possess more data will also enjoy higher market status and market power. Competition among capital still follows what Marx called the law of capital concentration; capital and data factors generally concentrate unidirectionally in the hands of enterprises that occupy favorable and dominant positions in the market, with large capital continuously swallowing up small capital. In the view of scholars such as K. Sabeel Rahman, under the platform model, the market competition structure moves toward oligopoly, and barriers to market entry will be built increasingly high.

第二，资本实行对数据要素的垄断。从表面上看，与资本、劳动、土地等传统生产要素不同，数据要素作为数字时代的重要生产要素具有不完全排他、不完全竞争等特点，但实际上并非如此。数据要素的来源十分丰富，人们在使用网络平台和社交媒体时，在不知不觉中为资本家贡献了个人数字劳动，生产了大量的数据。资本家可以近乎无偿地占有数字劳动者的劳动产品并侵吞剩余价值。祖博夫指出，无论是人们在网上的浏览记录、发布的内容，还是个人的身体数据信息，都能够被监视资本家暗中抓取并私自占有。掌握数据越多的资本家也将拥有更高的市场地位和市场权力。资本之间的竞争仍然服从马克思所说的资本集中规律，资本和数据要素总体上会单向集中到在市场上占据有利和主导地位的企业手中，大资本不断兼并小资本。在K.萨比尔·拉赫曼（K.Sabeel Rahman）等学者看来，在平台模式下，市场竞争结构向寡头垄断的方向迈进，市场的进入壁垒也将越筑越高。

Third, capital holds absolute power in the labor market. On platforms, intelligent algorithms reduce living workers to a series of strings; workers’ personal characteristics, preferences, and personalities are all symbolized. In the eyes of capitalists, the role of workers is only to provide homogenized labor power. In the past, if an enterprise wanted to hire employees in the labor market, it would typically conduct a comprehensive assessment of the candidate’s family background, educational background, knowledge structure, speech and behavior, etc., and after completing a series of procedures, finally sign an employment contract. Artificial intelligence, however, greatly simplifies the relevant processes. Capitalists do not overly concern themselves with the physical and mental health, emotions, or family conditions of workers hired through outsourcing and crowdsourcing platforms, because these workers only need to complete work tasks within the specified time. Gray argues that high-tech enterprises, through technologies such as big data, intelligent algorithms, and cloud computing, continuously outsource as many matters as possible—such as distributed databases and employee dispatch—to enhance their technological reputation or industry standing, while the original enterprise employees are forced to be laid off. Many startups in Silicon Valley in the United States choose to hire temporary workers or workers contracted through human resources agencies, which also allows the company to maintain a “lean” image on its books.

第三，资本在劳动力市场中掌握绝对权力。在平台上，智能算法将活生生的劳动者简化为一系列的字符串，劳动者的个人特征、爱好、性格都被符号化。在资本家的眼中，劳动者的作用只在于能够提供同质化的劳动力。过去，如果企业想要在劳动力市场中聘用员工，通常会对意向人选的家庭背景、学历背景、知识结构、言谈举止等进行综合考察，完成一系列程序后再最终签订雇佣合同。而人工智能则极大地简化了相关流程，资本家不会过度关心通过外包和众包平台雇佣到的劳动者的身心健康、情感、家庭等状况，因为这些工人只要能够在规定的时间内完成工作任务就可以了。格雷认为，高科技企业通过大数据、智能算法和云计算等技术，不断地将分布式数据库、员工派遣等尽可能多的事项外包出去，以提升自己的科技声誉或行业声望，原有的企业雇佣员工则被迫下岗。美国硅谷的许多初创企业都会选择雇佣临时工或通过人力资源机构签约的临时工来开展工作，这样还能使公司的账面保持“精益”的形象。

(II) The Exploitation of Labor by Capital Continues to Intensify

（二）资本对劳动的剥削持续加剧

First, workers’ working hours are forcibly extended, and labor compensation is depressed. Artificial intelligence enables capitalists to exploit workers even when they are not physically present, which objectively expands the scope of exploitation. In the context of intelligentization, the means of production still belong to capitalists, and the actual status of workers has not changed; they remain in a subordinate position. The development of technology has always served as a means for capital to extract surplus labor. Janine Berg, using the well-known employment platforms Amazon Mechanical Turk (MTurk) and CrowdFlower as case studies, discusses the impact of artificial intelligence on the working conditions and wages of ordinary workers. She points out that low pay is a common problem faced by platform workers. To earn more money, platform workers often work at night, and many even engage in unpaid work. Low-skilled workers are often in a high-pressure environment for long periods, not only with meager income and bleak prospects but also with heavy personal psychological burdens. The situation of algorithm engineers and programmers is also not optimistic. To secure job opportunities, ordinary programmers must make sustained efforts to participate in “bidding marathons” dominated by large enterprises, yet they cannot obtain guaranteed labor compensation.

第一，劳动者的工作时间被迫延长，劳动报酬被压低。人工智能使劳动者不必在场也仍然能够为资本家所剥削，这在客观上扩大了剥削的范围。在智能化的大背景下，生产资料仍然归资本家所有，劳动者的实际地位并没有改变，仍处于被支配的地位。技术的发展始终是服务于资本榨取剩余劳动的一种手段。珍妮·伯格（Janine Berg）以知名的雇工平台亚马逊土耳其机器人网站（Amazon Mechanical Turk, MTurk）和众花（CrowdFlower）为案例，讨论了人工智能对普通工人工作状态和薪酬的影响。她指出，薪酬过低是平台劳工面临的普遍问题。为赚到更多的钱，平台劳工在夜间工作是常态，甚至还有很多人从事着无酬的工作。低技能劳动者往往长期处在有着巨大压力的环境之中，不仅收入微薄、前景黯淡，个人心理负担也非常沉重。算法工程师和程序员的境遇也不容乐观。为争取到工作机会，普通的程序员不得不付出持久努力，参与到大型企业主导的“竞标马拉松”之中，但是他们却无法拿到确定的劳动报酬。

Second, workers’ autonomy is further reduced. Artificial intelligence either leaves people unemployed or forces them into industries with lower skill requirements. Unemployed workers often suffer from material deprivation, psychological emptiness, and low social status. According to reports, an employee of a U.S. company engaged in data mining, when discussing their work status, stated that they had no autonomy in arranging their work at all, because all tasks had to be completed on the platform; they could not choose their working hours, and once a task appeared, they had to complete it. Gray points out that if platform workers want to earn labor compensation, they must accept the platform’s irregular work arrangements and workflows. Algorithmic errors and malfunctions can also deeply trouble platform workers. For example, in 2013, a minor glitch in the Amazon Mechanical Turk system froze some registered worker accounts, causing these innocent workers to lose hundreds of dollars. Some refer to the harm caused by the unconscious behavior of artificial intelligence as “inadvertent algorithmic cruelty.” However, to some extent, it is not artificial intelligence itself that favors or harms specific groups, but rather the capitalist application of artificial intelligence that exerts a cruel impact on ordinary workers.

第二，劳动者的自主性进一步降低。人工智能让人们要么失业，要么被迫进入对劳动技能要求更低的行业。失业的劳动者通常会出现物质生活匮乏、心理空虚、社会地位低下等问题。据报道，美国一家从事数据挖掘业务的公司员工在谈及工作状态时表示，自己完全无法自主安排所做的工作，因为所有的任务必须在平台上完成，自己无法选择工作时间，一旦有任务就必须去完成。格雷指出，平台劳工若想赚取劳动报酬，就必须接受平台不规律的工作安排和工作流程。而算法程序如果出现错误和紊乱则会让平台劳工深受其扰。例如， 2013年，亚马逊土耳其机器人网站系统曾出现一个小故障，冻结了一部分登记注册的工人账号，导致这些无辜的劳动者损失了数百美元。有人将人工智能的无意识行为给人们带来的伤害称作“无意的算法残酷”（inadvertent algorithmic cruelty）。然而从某种程度而言，并非人工智能本身对特定群体进行了偏袒或伤害，而是人工智能的资本主义应用对普通劳动者造成了残酷的影响。

Third, the forms of exploitation suffered by workers have become more diverse. In the digital age, alongside physical and mental labour, new forms of work such as “playbor” and “affective labour” have emerged. The diversification of work forms means that the content of exploitation has expanded: while appropriating workers’ surplus labour without compensation, capitalists further appropriate workers’ thoughts and emotions, leisure time, personal information, social relationships, and so on, also without compensation. In addition, new exploitation is reflected in workers’ inability to enjoy the protection of labour policies. In traditional production organisations, capitalists must pay workers a minimum wage and guarantee corresponding welfare benefits. However, with the platform-based transformation of production organisations, platform workers may have their wages arbitrarily reduced or be subjected to exorbitant fines. Research by Rahman and others shows that in reality many companies no longer sign labour contracts, thereby avoiding the need to bear expenditures on workers’ wages, welfare, and insurance. This allows capitalists simultaneously to achieve three objectives: reducing labour costs, evading regulatory oversight, and strengthening control over the production network.

第三，劳动者遭受剥削的形式更加多元。在数字时代，在体力劳动、脑力劳动之外，又出现了“玩劳动”（playbor）、“情感劳动”（affective labor）等新的工作形式。工作形式的多元化意味着剥削内容发生了延展，资本家在无偿占有劳动者剩余劳动的同时，还会进一步无偿占有劳动者的思想情感、娱乐时间、个人信息、社交关系等。此外，新的剥削还体现在劳动者无法享受劳动保护政策的庇护等方面。在传统的生产组织中，资本家须向劳动者支付最低工资、保障相应的福利待遇。但是，随着生产组织的平台化变迁，平台工人会被肆意压低工资或是被判处高额罚金。拉赫曼等人的研究显示，现实中许多公司不再签订劳动合同，从而也就无需承担劳动者的工资、福利和保险等方面的支出。这便让资本家同时实现了削减劳动力成本、规避执法监督和加强对生产制造网络的控制等三个目的。

(3) The contradiction between labour and capital becomes concealed

（三）劳动与资本的矛盾走向隐匿

The mode of combining labour and capital, with artificial intelligence as an important carrier, makes the contradictory relationship between labour and capital more hidden. As Gray and Suri put it: “What actually drives many mobile phone applications, websites, and AI systems is human labour, but this is hard to detect – in fact, this labour force is often deliberately hidden.”

以人工智能为重要载体的劳资结合方式使劳动与资本之间的矛盾关系变得更加隐蔽。正如格雷和苏里所说的：“真正驱动许多手机应用程序、网站和人工智能系统运行的竟是人类的劳动力，但我们很难发现——事实上，这些劳动力往往是被故意隐藏起来的。”

On the one hand, the production process of surplus value becomes concealed. Marx once pointed out: “The time the worker works beyond the limits of necessary labour, although it costs the worker labour, costs labour-power, yet does not form any value for the worker. This time forms surplus value, which attracts the capitalist with all the charm of creation out of nothing.” [4]4 Karl Marx, *Capital*, Vol. 1, Chapter 8, Section 1 (Penguin edition, p. 325). Artificial intelligence provides a new stable site for capital to exploit surplus labour. Capitalists crowdsource work tasks to network users; these users, while being consumers, are also workers, and they work for capitalists at extremely low remuneration. Trebor Scholz points out that people usually treat online platforms as playgrounds and do not realise they are actually factories. While traditional labour still exists, some unpaid informal labour has emerged in society; whether in form or content, this labour does not appear to be labour. Capitalists cleverly conceal themselves behind technology, setting up various unobservable barriers to maximise their own interests. Compared with the past, workers in the age of artificial intelligence are more likely to accept this form of work willingly, but the essence of capital’s exploitation of labour has not changed. Advanced technology has brought capital’s exploitation and alienation of labour into a new stage, and it is very difficult for workers to confront capitalists directly.

一方面是剩余价值的生产过程变得隐蔽。马克思曾指出：“工人超出必要劳动的界限做工的时间，虽然耗费工人的劳动，耗费劳动力，但并不为工人形成任何价值。这段时间形成剩余价值，剩余价值以从无生有的全部魅力引诱着资本家。”人工智能为资本剥削剩余劳动提供了新的稳定场所。资本家将工作任务众包给网络使用者，这些使用者作为消费者的同时又是劳动者，他们在劳动报酬极低的情形下，为资本家卖命工作。特雷博·肖尔茨（Trebor Scholz）就指出，人们通常把在线平台当作是游乐场，并未发现它其实是工厂。在传统劳动仍然存在的同时，社会中出现了一些无偿的非正规劳动，无论是从形式还是内容上看，这些劳动都不像是劳动。资本家巧妙地将自己隐匿于技术背后，设置各种不可观测的壁垒以实现自身利益最大化。与过去相比，人工智能时代的劳动者更有可能愉快地接受这种工作形式，但资本剥削劳动的本质并没有改变。先进技术将资本对劳动的剥削和异化带入了一个新的阶段，劳动者很难有机会直接面对资本家。

On the other hand, capital’s control over labour becomes concealed. Capitalists, with the help of artificial intelligence, have strengthened their control over workers while simultaneously making the form of control hidden. Scholars such as Andersson Schwarz have pointed out in their research that capitalists, when operating online platforms, set up various unequal terms and agreements in order to carry out unreasonable surveillance of consumers and strengthen real-time supervision of workers. These practices are not easily detected and are key links in capital’s extraction of surplus value. Moreover, the normal operation of online platforms requires a certain number of employers and a vast number of workers; a large amount of capital and anonymous workers form a two-sided market. In this two-sided market, platform capitalists claim to be “neutral”, not favouring either side, merely providing a matching mechanism to eliminate information barriers and prudently managing the “two-sided market”. But in fact, the algorithmic mechanism behind the platform serves entirely the fundamental purpose of platform capitalists’ extraction of surplus value; capitalists manipulate workers and extract surplus labour through the “algorithmic black box”. [5]5 “Algorithmic black box” refers to the opaque, non-transparent operation of algorithms, whose inner workings and decision-making criteria are hidden from users and workers. The result is that even if workers feel some kind of oppression, because there is no fixed workplace or working hours and they are unclear about who employs them, they directly attribute this oppression to the involutionary competition [6]6 “Involutionary competition” (内卷化竞争, *neijuanhua jingzheng*) describes a situation of intense, self-defeating competition within a closed system that yields no genuine progress or qualitative improvement. of other workers, the non-benevolent behaviour of consumers, or the intelligent algorithm itself, and spend more time and energy searching for and waiting for work opportunities.

另一方面是资本对劳动的控制变得隐蔽。资本家借助人工智能加强了对劳动者的控制，同时又使控制形式变得隐蔽。安德森·施瓦茨（Andersson Schwarz）等学者在研究中指出，资本家在运营在线平台时设置了各种不平等条款与协议，以便开展对消费者的不合理监控活动并加强对劳动者的实时监督。这些行为不易被发现，且都是资本攫取剩余价值的关键环节。此外，在线平台的正常运转需要一定数量的雇主和数量庞大的劳动者，大量资本和匿名劳动者形成了双边市场。在这种双边市场中，平台资本家辩称自己是“中立”的，不会偏向任何一方，只是为了消除信息壁垒而提供一种匹配机制，审慎管理着“双边市场”。但事实上，平台背后的算法机制完全服务于平台资本家获取剩余价值的根本目的，资本家通过“算法黑箱”来操控劳动者并榨取剩余劳动。结果便是，劳动者即使感受到了某种压迫，也会因为没有固定的工作场所、工作时间，不清楚受雇于谁，而直接将这种压迫归因于其他劳动者的内卷化竞争、消费者的非善意行为或是智能算法本身，并花费更多的时间和精力去搜寻和等待工作机会。

IV. Conclusion

四、结语

The research in this article shows that, driven by the new round of technological revolution led by artificial intelligence, the capitalist mode of production has undergone profound changes. On the one hand, the application of artificial intelligence has profoundly transformed the system of machinery, promoting the improvement of social productive forces and social production efficiency; on the other hand, artificial intelligence is accelerating its fusion with workers’ living labour and beginning to exist as a powerful production mechanism and mode of social operation. Marx emphasised in Volume 1 of *Capital*: “The capitalist mode of production and the development of the productive power of labour – both the cause and the effect of accumulation – enable the capitalist, by extending the scale of exploitation of individual labour-power, both extensively and intensively, to exploit it more.” The reality of capitalist development confirms this important thesis of Marx. Under the dominance of the logic of capital, artificial intelligence has profoundly reshaped the forms of production organisation, employment methods, and labour patterns; capital’s control and exploitation of labour have developed to a deeper stage. Capitalists, using new technologies and new means, have significantly increased the intensity of exploitation of surplus labour. The capitalist application of artificial intelligence once again intensifies the contradiction between the bourgeoisie and the proletariat, and brews a comprehensive crisis within capitalist society.

本文的研究表明，在人工智能引领的新一轮科技革命的推动下，资本主义生产方式发生了深刻变迁。一方面，人工智能的应用深刻地改造了机器体系，促进了社会生产力和社会生产效率的提高；另一方面，人工智能与工人的活劳动加速融合，并开始作为一种强大的生产机制和社会运转方式而存在。马克思在《资本论》第1卷中强调：“资本主义生产方式和劳动生产力的发展——既是积累的原因，又是积累的结果——使资本家能够通过从外延方面或内涵方面加强对单个劳动力的剥削。”资本主义发展的现实证实了马克思的这一重要论断。在资本逻辑的主导下，人工智能深刻地重塑了生产组织形式、雇佣方式、劳动形态等各个方面，资本对劳动的控制与剥削发展到了一个更为深化的阶段，资本家借助于新技术和新手段显著加大了对剩余劳动的剥削力度。人工智能的资本主义应用再次激化了资产阶级和工人阶级之间的矛盾，并在资本主义社会内部酝酿着全面危机。

(Author: Wang Wenze, School of Economics, Renmin University of China)

（作者：王文泽，中国人民大学经济学院）

Online Editor: Tongxin

网络编辑：同心