A humanoid robot playing an electronic keyboard at the 2025 GBA International Embodied Intelligence Robot Industry Conference and Exhibition, held in Guangzhou, Guangdong Province, from Oct. 22 to 24 Photo: IC PHOTO

Embodied intelligence represents a new frontier and direction in artificial intelligence (AI). While propelling the advancement of AI technologies, it also opens new possibilities for human–machine interaction and integration. As it profoundly reshapes contemporary society and drives the ongoing evolution of the intelligent age, it likewise raises a series of new philosophical questions, broadening the horizons of philosophical inquiry and prompting forward-looking reflection.

From embodied cognition to embodied intelligence

Research on embodied intelligence originated from reflections on traditional cognitive science, particularly critiques of mind–body dualism. It emphasizes the significance of the body and environment in cognitive processes, arguing that human cognition is deeply rooted in bodily structures and lived experience. Our sensory perception, motor abilities, emotional responses, and sociocultural context collectively shape how we think and understand the world.

For example, visual perception depends not only on the physiological structure of the eyes but also on how we move our bodies and heads, and how we use our hands to explore our surroundings. Thus, cognition is not merely a function of the brain but a product of the body’s interaction with the world—a process through which the body engages with its environment and gains understanding.

Embodied intelligence asserts that intelligence itself is embodied, based on the premise that the emergence and development of intelligence are closely tied to perception, action, and modes of interaction with the environment. In other words, intelligence is not simply information processing within the brain but an integrated process shaped by bodily structure and reciprocal engagement with the external world. This perspective underscores the decisive role of the body in cognition, affirming that intelligence and cognition cannot be understood apart from their concrete physical existence. Building on this foundation, embodied intelligence seeks to realize more natural and adaptable intelligent behaviors through dynamic interactions between body and environment.

Since the theory of embodied cognition preceded that of embodied intelligence, the latter can be seen as its extension, projecting the embodied stance of human cognition into the domain of artificial intelligence and revealing the embodied nature of AI itself. Embodied intelligence achieves intelligent behavior through physical entities that perceive, move, and interact with their surroundings—mirroring how bodily actions, sensations, and motor experiences shape cognition in humans. It aims to generate extended intelligence through extended bodies, or to cultivate machine intelligence through machine embodiment.

An intelligent agent’s body and sensors actively participate in the formation of experience, perceiving the external world to generate experiential data rather than merely receiving preprocessed, high-quality data as input. In this sense, the rise of embodied cognition has spurred a new understanding of intelligent behavior within AI. The theories and methods of embodied cognition have inspired research on embodied intelligence, driving its development and enabling AI experts to design intelligent machines capable of effective interaction with the real world.

Toward ‘fused intelligence’

When embodied cognition and embodied intelligence converge through human–machine integration, they give rise to a new paradigm—a fusionist view of embodied cognition. This perspective proposes that a fused body—composed of human and AI elements—can generate a new, hybrid form of intelligence that not only inherits human creativity but also incorporates AI’s computational speed and data-processing power.

More importantly, this intelligence is embodied: It depends on the physical characteristics of the fused body and its interaction with the environment for its development and expression. Fused intelligence does not exist in abstraction; it is deeply embedded in every aspect of the fused body and its engagement with the world. The cognitive process no longer belongs to a single entity because it transcends the boundary between human and machine. Cognition or intelligence can thus emerge only from a fused body, and fused intelligence can only exist within such embodiments.

This fusionist perspective expands and redefines the horizons of embodied cognition, offering new theoretical support. It shows that the fusion of material forms enables the fusion of information. In terms of the body–intelligence configuration, intelligence can be integrated only through bodily integration in the broad sense, since intelligence always requires embodiment—there is no such thing as “bare intelligence” detached from any physical substrate. On this new embodied foundation, the inseparability of intelligence and the body is reaffirmed.

Embodied intelligence also reminds us that the limits of intelligence largely stem from the limitations of the body. For human intelligence, the finitude of the body necessarily constrains the level of intelligence. From a broader perspective, only a fused body can produce fused intelligence; “higher” or “more advanced” forms of intelligence must rest upon newer, more integrated bodies. In this sense, the evolution of intelligence depends on the evolution of embodiment. Humanity’s qualitative leap in intelligence can be realized only through the integration of the body—especially the human brain—with embodied intelligence, that is, through a bodily revolution leading to an intelligence breakthrough.

When it comes to the ownership of embodied intelligence resulting from human–machine integration, several questions arise. Is it a new form of artificial intelligence, a redefined form of human intelligence, or perhaps a third kind distinct from both? Such questions invite broader reflection on the diversification of intelligence. Furthermore, if intelligence can be fused, can mind be fused as well? If a fused body and fused brain can generate fused intelligence, might they also produce a fused “human–machine mind?” Would such a fused mind possess a hybrid self-consciousness distinct from that of humans?

These reflections suggest that embodied intelligence not only propels the technological progress of AI but also deepens our philosophical reconsideration of intelligence, mind, and embodiment—potentially guiding us toward more advanced and flexible intelligent system designs.

Humanity in age of human-machine integration

As human–machine integration deepens, future societies may witness a more harmonious, symbiotic relationship between humans and intelligent machines based on complementarity. Machines will no longer be mere tools but partners capable of participating in decision-making, creation, and even assuming social roles. When machines act as autonomous agents in social life, their errors or malicious actions could have serious consequences, requiring the establishment of ethical frameworks and regulatory mechanisms for machines. In the context of deep human-machine integration, when mistakes or harm occur, how should responsibility be assigned? Should it lie with the designer’s negligence, the user’s improper operation, or the “intentions” of the machine itself?

As machine intelligence grows and autonomy increases, machines shift from tools to potential collaborators—or even competitors. This calls for a reassessment of their role and value in human society, including the power dynamics and social norms between humans and machines, and especially how humans define and preserve their own value and dignity in an increasingly intelligent world. For example, human–machine integration may challenge conventional notions of human essence: Are intelligence, consciousness, and emotion uniquely human attributes, and how do these traits define human distinctiveness?

Artificial intelligence, especially embodied intelligence, is accelerating transformations of human “bodily existence,” pushing humanity toward a “technologized body.” Such transformations raise new questions about self-identity: An individual’s identity may no longer be confined to the biological body but may include technological augmentations. This inevitably sparks philosophical debates about “what it means to be human.” If a cybernetic body equipped with cyber intelligence becomes feasible, a new species could emerge. Whether this species should still be called “human” remains a subject of intense debate, with bioconservatism and posthumanism representing two opposing views.

Bioconservatives argue that true “humans” should retain their natural biological characteristics. They worry that excessive reliance on technology may cause humanity to lose essential traits such as emotion, creativity, and moral judgment. They generally oppose any form of genetic editing, cognitive enhancement, and bodily modification, fearing that such technologies could undermine human identity and social structures. They also caution that dependence on machines could create ethical confusion, raising questions about whether machines should possess rights and how to delineate boundaries between humans and machines.

Posthumanists, by contrast, take a more open stance, viewing technology as a means for further human evolution. They argue that enhancing human physical and cognitive capacities through technology can overcome biological limitations and achieve higher levels of intelligence, health, and longevity. Posthumanists typically support the research and development of technologies that augment human capabilities—such as genetic editing, neural implants, and bionic limbs—seeing these developments not only as possible but as necessary to address major challenges facing humanity, including disease and aging.

Although bioconservatism and posthumanism are often presented as opposites, some degree of synthesis is possible. For instance, some may accept limited technological enhancements, such as gene therapy for medical treatment, while rejecting radical biological remodeling. Moreover, as technology advances, our understanding of what it means to be “human” continues to evolve, giving rise to perspectives that fall between the two extremes. Ultimately, the debate over whether a “cybernetic body with cyber intelligence” still counts as human is, at its core, a discussion of the relationship between technological progress and human identity.

If embodied intelligence represents the extension of artificial intelligence into the body, is integrative intelligence humanity’s “domestication” of embodied intelligence—or an “if you can’t beat them, join them” strategy? Could these two positions be reconciled to achieve a balanced coexistence? Naturally, avoiding mutual exhaustion, conflict, or destruction between humans and machines will remain a constant concern as human society enters the era of embodied intelligence.

In sum, human–machine integration within embodied intelligence ushers in an era filled with both opportunities and challenges. It demands not only continuous technological innovation but also profound reflection and adaptation in ethical, legal, and social domains to ensure that technological progress continues to serve human well-being while safeguarding human value and dignity.

Xiao Feng is a professor from the School of Marxism at Shanghai University. This article has been edited and excerpted from Hebei Academic Journal, Issue 3, 2025.