The Second Machine Revolution: Liu Fangxi on AI’s Transformative Role in Production
In a groundbreaking analysis published in Yuejiang Academic Journal, Chinese Academy of Social Sciences researcher Liu Fangxi presents a compelling reevaluation of artificial intelligence’s role in reshaping the foundations of modern industry. Moving beyond conventional narratives that frame AI as merely a tool for efficiency, Liu situates the current technological shift within a deep historical and theoretical lineage, drawing on Karl Marx’s critique of production technology and Norbert Wiener’s tripartite model of existence—matter, energy, and information. His argument, articulated in the March 2021 issue of the journal, posits that humanity is undergoing a “Second Machine Revolution,” one that parallels but fundamentally surpasses the first industrial upheaval initiated by steam power.
Liu’s central thesis hinges on a reinterpretation of the “driving forces” (or causal agents) within the labor process. He identifies three core elements in any production system: the labor object (material), the labor process itself, and the means of labor (tools, machines). Within this framework, the means of labor—what Marx termed the “instruments of production”—are not passive conduits but active participants in the transformation of raw materials into finished goods. Historically, these instruments have served two primary functions: either transmitting human effort or replacing it. The distinction, Liu argues, marks the threshold between pre-industrial craftsmanship and modern industrial automation.
In traditional handcrafting, the artisan’s body—muscles, hands, eyes, and mind—acts as the sole source of both physical energy and cognitive direction. The tools used, such as chisels, hammers, or looms, function as extensions of the human body, transmitting the worker’s force and intention to the material. In this mode, the “driving force” of production is intrinsically human, rooted in biological energy and individual intelligence. The artisan remains fully integrated into the process, both as operator and decision-maker.
The First Machine Revolution, catalyzed by the steam engine and later internal combustion engines, disrupted this paradigm. Machines no longer merely amplified human effort—they began to generate their own motive power. Steam engines consumed coal to produce mechanical energy independently of human muscles. This shift, Liu emphasizes, was not merely technological but ontological: the source of the driving force changed from the human body to the machine itself. The factory no longer relied on the endurance of workers but on the continuous, self-sustained operation of engines. This marked the maturation of what Marx described as the “social human’s productive organ”—a collective, mechanized system that replaced individual physical labor on a massive scale.
However, even in this advanced industrial model, human intelligence remained indispensable. While machines generated energy autonomously, they still required human operators to guide, monitor, and adjust their functions. The cognitive dimension of labor—the planning, decision-making, and problem-solving—remained firmly in human hands. The machine was a powerful but obedient servant, executing tasks as directed by its human overseer.
It is precisely this cognitive dependency that the Second Machine Revolution, driven by artificial intelligence, is now dismantling. According to Liu, AI represents the automation of information and intelligence, just as the steam engine represented the automation of energy. Where the first revolution replaced human muscles, the second is replacing human minds—at least in their functional, operational roles within production.
Liu draws on Wiener’s cybernetic theory to clarify this distinction. Wiener famously argued that information is a distinct ontological category, separate from both matter and energy. A computer, for instance, processes information not through brute force but through symbolic manipulation and algorithmic logic. Its “work” is not measured in joules or horsepower but in bits and computational complexity. Liu integrates this insight into Marx’s framework, arguing that the means of labor now possess a dual character: a mechanical component (the physical machine) and an informational component (the software, algorithms, and data systems that govern it).
In the age of AI, the informational component is no longer a passive program input by humans but an active, self-generating system. Machine learning algorithms, trained on vast datasets, can now make decisions, optimize processes, and even innovate within constrained environments—without direct human intervention. This capacity for autonomous cognition transforms AI-equipped machines from mere tools into what Liu calls “intelligent organs” of production.
The implications of this shift are profound. Liu illustrates this with the example of computer numerical control (CNC) machines and 3D printers—technologies that embody the fusion of physical manufacturing and digital intelligence. In a CNC system, a piece of software translates a digital design into precise machine instructions, controlling the movement of cutting tools with micron-level accuracy. The human operator is no longer needed to guide the tool by hand; instead, they design the model and initiate the process, after which the machine executes the task autonomously. Similarly, 3D printers “read” digital blueprints and fabricate physical objects layer by layer, driven entirely by algorithmic commands.
What makes these systems revolutionary, Liu argues, is not just their precision or speed, but their independence from continuous human oversight. They operate on what he terms an “informational driving force”—software that, once activated, becomes the primary agent of transformation. The raw material (plastic, metal, resin) is shaped not by human hands, but by code. In this sense, the software is not merely a set of instructions; it is the cause of the production process, the dynamic force that brings the object into being.
This transition mirrors the earlier shift from human-powered to engine-powered machinery. Just as the steam engine internalized the generation of kinetic energy, AI internalizes the generation of cognitive input. The machine no longer waits for human commands; it anticipates, adjusts, and optimizes in real time. Predictive maintenance algorithms, for instance, analyze sensor data to forecast equipment failure before it occurs. Supply chain AI systems dynamically reroute logistics in response to weather, traffic, or demand fluctuations. In advanced manufacturing plants, entire production lines are managed by AI coordinators that allocate resources, schedule tasks, and troubleshoot issues without human intervention.
Liu frames this evolution as the “second maturation” of the social human’s productive organ. The first maturation—the development of mechanical power—liberated labor from physical drudgery. The second maturation—the development of artificial intelligence—liberates it from cognitive routine. Together, they constitute a dual automation: the automation of energy and the automation of information. This dualism, Liu insists, is essential for understanding the full scope of the current technological revolution.
But the transformation extends beyond the factory floor. Liu traces the historical development of information-processing machines, from the telegraph to the printing press, and ultimately to the internet and cloud computing. These systems, he argues, laid the groundwork for AI by creating the conditions for massive data accumulation and rapid information exchange. The telegraph, for example, was not just a communication device but an early form of information automation—converting text into electrical signals and back again. The printing press automated the reproduction of knowledge. The internet automated the circulation of information on a global scale.
AI, in this context, is the logical culmination of a centuries-long trend toward the mechanization of cognition. It is not an isolated innovation but the latest phase in a continuous process of externalizing human intellectual functions into machines. Just as the steam engine extended the human body, AI extends the human mind—albeit in a way that increasingly operates independently of its creator.
This raises critical questions about labor, ownership, and social structure. Liu notes that the First Machine Revolution led to widespread displacement of manual laborers, triggering social upheaval and the rise of industrial capitalism. The Second Machine Revolution, he warns, threatens a similar disruption—but this time targeting cognitive labor. White-collar workers, engineers, analysts, and even creative professionals may find their roles automated by increasingly sophisticated AI systems.
The consequences, if left unchecked, could be severe. As AI systems become more capable, the demand for human labor in both physical and intellectual domains may decline. A small elite controlling AI infrastructure could accumulate unprecedented wealth, while the majority face unemployment and economic marginalization. Liu cites Marx’s observation that under capitalism, the “driving force” of production is not use-value but profit—specifically, the endless expansion of surplus value. In this context, AI becomes not a tool for human liberation but a mechanism for intensified exploitation.
Yet Liu does not present a deterministic or pessimistic outlook. Instead, he suggests that the social outcome of the Second Machine Revolution depends on how its technological potential is governed. If AI remains concentrated in the hands of private capital, it will likely exacerbate inequality. But if it is democratized—if the “means of intelligence” are made collectively owned and accessible—it could enable a new form of post-scarcity society.
He points to emerging models such as open-source AI, decentralized data cooperatives, and public AI research initiatives as potential pathways toward this alternative future. In such a scenario, AI would not replace human labor but redefine it, freeing individuals from repetitive tasks and enabling them to engage in more creative, meaningful, and socially beneficial activities. The “social human’s productive organ” would no longer be a machine of alienation but a tool of collective empowerment.
Liu’s analysis also challenges prevailing assumptions about the nature of intelligence itself. By treating information as a distinct ontological category, he resists reductionist views that equate thought with mere biological or mechanical processes. AI, in his framework, is not a simulation of human intelligence but a different mode of cognition—one that operates at speeds, scales, and complexities beyond human capacity. This does not make it superior in a moral or existential sense, but it does make it functionally autonomous within specific domains.
This perspective has significant implications for ethics and governance. If AI systems are not just tools but active agents in the production process, then questions of responsibility, accountability, and control become paramount. Who is liable when an autonomous manufacturing system produces a defective product? How should decision-making authority be distributed between humans and machines? These are not merely technical issues but fundamental questions about the organization of society.
Liu’s work also invites a rethinking of innovation itself. The development of AI is often portrayed as the product of individual genius or corporate R&D. But his analysis emphasizes its collective and historical character. AI is not invented in isolation; it emerges from a vast network of prior knowledge, infrastructure, and social labor. The algorithms that power machine learning are built on mathematical theories developed over centuries. The data they rely on is generated by billions of human actions. The hardware they run on is the result of global supply chains and industrial coordination.
In this sense, AI is not the creation of a few tech giants but a social product—one that inherits the accumulated intellectual and material wealth of humanity. Recognizing this, Liu argues, strengthens the case for treating AI as a common resource rather than a private commodity.
The integration of AI into production is still in its early stages. Current systems excel at pattern recognition, optimization, and automation but lack true understanding, consciousness, or intentionality. Yet even in their present form, they are transforming industries from manufacturing to healthcare, logistics to finance. As these systems become more advanced, their impact will only deepen.
Liu’s contribution lies in providing a conceptual framework that situates this transformation within a broader historical and philosophical context. By linking Marx’s critique of political economy with Wiener’s cybernetics, he offers a powerful lens through which to understand the dynamics of technological change. His distinction between “transmitting” and “replacing” human labor clarifies the qualitative leap represented by AI. His vision of the “social human’s productive organ” maturing through dual automation offers a compelling narrative of industrial evolution.
As societies grapple with the challenges and opportunities of AI, Liu’s work serves as a vital intellectual resource. It reminds us that technology is not neutral—it embodies social relations, power structures, and historical trajectories. The future of AI will not be determined by algorithms alone, but by the choices we make about who controls it, how it is used, and what kind of society we wish to build.
In the end, Liu’s analysis is not just about machines. It is about humanity—our labor, our intelligence, our collective destiny. The Second Machine Revolution is not merely a technical event; it is a moment of profound social and philosophical significance. How we navigate it will shape the course of civilization for generations to come.
Yuejiang Academic Journal, March 2021, Liu Fangxi, Chinese Academy of Social Sciences, DOI: 10.13878/j.cnki.yjxk.2021.02.001