In the summer of 2009, neuroscientist Anil Seth encountered a dozen common octopuses (*Octopus vulgaris*) at a lab in Naples. These creatures, capable of dramatically altering their shape, color, and texture, captivated Seth. He observed them paying intense attention to the work of his colleagues, seemingly driven by pure curiosity. This experience profoundly shifted Seth's perspective on consciousness, suggesting that non-human minds could be vastly different from our own. Octopus consciousness, he postulates, might be more distributed and less integrated than human consciousness, potentially lacking a central control point. With octopuses able to detect light through their skin, it's even conceivable their central brain is unaware of these peripheral sensory inputs. Furthermore, their arms exhibit semi-autonomous behavior, leading to a less stable sense of body ownership compared to humans. Seth famously described the octopus mind as "an independently created evolutionary experiment, as close to the mind of an alien as we are likely to encounter on this planet."
Despite their 'alien' nature, Seth views octopuses as genuine biological kin, a connection he fears is overshadowed by our fascination with artificial intelligence. As a professor of Cognitive and Computational Neuroscience, Seth has extensively pondered how humans have come to equate themselves with AI systems. He posits a "two-way mirror" effect, where we perceive ourselves through the technologies we develop. The long-standing academic metaphor of the brain as a computer has gained new traction with the advent of sophisticated AI, leading some to conclude that humans are merely complex computational systems. This perspective is further fueled by discussions around large language models, sometimes dismissed as "stochastic parrots" – systems that generate human-like text based on statistical probabilities without genuine comprehension. However, Seth cautions that this reductionist view, often termed "mechanizing our minds," is diminishing and fails to capture the full essence of human experience, which is deeply rooted in lived experience, consciousness, and our biological makeup.
The Shifting Metaphor: From Organism to Algorithm
Seth frames the tendency to equate humans with machines as a continuation of an age-old human endeavor: understanding our place within nature. When faced with phenomena that resist easy explanation, humans resort to metaphors. The brain, in particular, has been a subject of intense metaphorical exploration. While calling the heart a pump is a relatively apt comparison, the brain-as-computer metaphor carries more significant implications. This analogy, normalized by decades of academic and technological discourse, implies a deeper congruence between biological brains and digital computers than may exist. Early animist cultures, for instance, perceived spirits animating both nature and humans, fostering a sense of parity. In contrast, our technological era increasingly looks to software or algorithms as the hidden "spirit" animating humans, thereby extending this logic to computers and suggesting their code grants them a form of agency beyond mere hardware.
This mechanistic view draws upon a historical human exceptionalism, which identifies the mind, language, and intelligence as uniquely human attributes, setting us apart from other animals and elevating us towards the divine. René Descartes famously articulated this by positing the mind as a disembodied thinking substance, capable of existing independently of the body. This philosophical stance arguably makes artificial intelligence, particularly language models, seem seductively familiar. Seth suggests that we resonate more with language models than with protein-folding AIs like AlphaFold because the former echo the very capacities long considered hallmarks of humanity. Seth himself has critically stated, "We're not cognitive computers, we're feeling machines." However, he readily acknowledges this is a "false opposition," citing the work of neuroscientists like Antonio Damasio, who demonstrate that emotion is integral to cognition. Effective reasoning, Seth argues, relies on the "feeling part of our humanity or of our animal nature"; without bodily input and emotion, sound decision-making is compromised.
The Abstraction of Biology
In recent decades, a decline in human exceptionalism has become apparent, marked by a growing recognition of consciousness in non-human animals and a better understanding of humanity's biological interconnectedness with nature. Yet, the older exceptionalist narrative persists, now reframed through the language of computation. Seth attributes this resurgence to two pivotal developments from approximately ninety years ago. Firstly, Alan Turing's conceptualization of computation as medium-independent—where algorithms map symbols to symbols, with the physical substrate only relevant for its ability to execute the algorithm—laid crucial groundwork. Secondly, Warren McCulloch and Walter Pitts demonstrated that simplified neuronal models, devoid of intricate biological detail, could implement Turing computations. According to Seth, these ideas formed a "mathematical marriage of convenience," enabling the dismissal of "messy biological detail."
This abstraction proved instrumental in the development of modern AI. However, it also fostered a pervasive belief: if the algorithm is all that matters, then the brain's metabolism, chemistry, and living texture can be dismissed as mere "implementation detail." Seth contends that this is a critical error. Human cognition is intrinsically linked to physical, continuous time, whereas algorithms primarily concern sequence and order. Unlike computers, which can separate hardware from software, biological brains cannot divorce "mindware" from "wetware." Real neurons fire due to complex metabolic, chemical, and biological imperatives that silicon substrates cannot replicate. Consequently, the assumption that brain function is independent of its biological substrate is flawed. Once this assumption is challenged, the notion that computation is the sole defining characteristic of the brain loses its foundation.
Understanding the Mirror: AI as a Catalyst for Self-Discovery
Seth clarifies that his critique of AI is not a call for disengagement. Instead, he sees "a lot of potential for synergy," where AI can serve as a tool for enhanced self-understanding, paradoxically highlighting our unique differences. The crucial question then evolves: "What is it that we do that is different still?" This remains an open inquiry, one that cannot be answered by fixating on superficial human traits. The search should focus on identifying non-trivial distinctions, acknowledging that while specific capacities might eventually be replicated by AI, human distinctiveness may lie in the intricate interplay of all properties that constitute a human being. Language, once considered a definitive human marker, now presents a complex case. While AI has demonstrated mastery in games like chess and Go, these are not universal human benchmarks. Language, however, is a near-universal human capacity, and the fluency of language models challenges this perceived demarcation between humans, animals, and technology.
Simultaneously, AI is aiding in deciphering animal communication, such as the utterances of dolphins, revealing a richness previously underestimated. This dual pressure—machines exhibiting linguistic capabilities and animals demonstrating complex communication—blurs established boundaries. Furthermore, the very definition of "understanding" is being questioned. While few attribute consciousness to current AI, many assert their capacity for understanding. Seth posits that understanding can occur unconsciously, suggesting that current language models might possess syntax without semantics. However, he contemplates that an embodied, embedded language model trained through real-world interaction could be deemed to truly understand, even without conscious experience. This challenges the anthropocentric assumption that understanding and consciousness are invariably linked, suggesting another potential "echo of anthropocentrism" in our definitions.
Beyond Language: The Embodied Nature of Consciousness
Consciousness itself remains a profound area of inquiry. Seth agrees that current AI systems lack consciousness, and many capabilities that living beings exercise consciously are still elusive for AI. A key difference lies in human learning: we acquire language from limited examples, generalize rapidly, use minimal energy, self-correct, and possess an intuitive sense of uncertainty. Seth attributes this efficiency to our embodied and embedded nature from the outset. While AI might not require consciousness to achieve similar feats, it suggests that consciousness in biology may confer functional advantages that are not yet fully understood. Seth poses the fundamental question: "What are the functions of consciousness? What does it do for us?"
The temporal dimension of human cognition also diverges significantly from algorithmic processes. Human consciousness unfolds within real, continuous, physical time, unlike algorithms that can theoretically run indefinitely. Biological beings are subject to imperatives like thirst and hunger, compelling action and adaptation within a flowing temporal landscape. This inherent need to act, Seth suggests, may assist in resolving the "frame problem"—the challenge of discerning relevant environmental features in a complex world. By acting, we alter the environment, break deadlocks, and progress. While AI can be programmed to operate in continuous time, the critical role of this temporal flow in human cognition is often overlooked when comparing brains to algorithms. Embodiment is another crucial differentiator: "We operate in the world. We don’t operate in a world of labels only." Much of cognition arises from our physical interactions with the environment, rather than solely from abstract computations within a disembodied brain. As Seth emphasizes, "We experience the world around us and ourselves within it—with, through and because of our living bodies." Over-reliance on substrate-independent computation as a sole explanation for consciousness risks narrowing our understanding and potentially over-attributing subjective experience to systems that may not possess it.
The Primacy of Embodiment and Feeling
Seth's perspective does not negate the possibility of consciousness in non-biological systems, but he advocates for caution. He highlights the assumption that consciousness is exclusive to living systems, despite the fact that every recognized instance of consciousness is biological. His core recommendation is humility: acknowledging our underlying assumptions, resisting a singular narrative of AI's path to consciousness, and fostering a broader inquiry into the nature of mind. Key questions include identifying what truly distinguishes the human mind, the extent to which brain function is dependent on its biological substrate, and how much can be abstracted into algorithmic processes.
The current AI landscape is prompting these questions with unprecedented urgency. Ideas once confined to philosophical discourse are now confronting practical implications, making philosophy increasingly relevant. The fundamental question of consciousness itself has gained a prominence it lacked even five years ago. Seth explores this through Thomas Metzinger's concept of "minimal phenomenal experience"—the most basic form of conscious experience a living organism could possess. While some theories propose pure awareness devoid of content, Seth considers the "feeling of being alive"—a fundamental, formless sensation—as a potential candidate. He hypothesizes that this core feeling might be essential for consciousness, suggesting that "it is life, rather than information processing, that breathes fire into the equations of experience." While acknowledging the speculative nature of this hypothesis, he finds it valuable for proposing a fundamental form of consciousness. He advocates for protecting and cultivating this essential feeling, which anchors us in our embodied existence rather than abstract thought, serving as an antidote to the "over-intellectualization, rationalization, cognitivization, algorithmicization, computationalization of the mind."
This embodied feeling offers an alternative to the dominant Western concept of the soul as an immaterial essence, separate from the body—a view that mirrors contemporary notions of "digital immortality" or mind uploading. Ancient traditions, such as the Greek *psychē* linked to breath and the Hindu *Ātman* representing witnessing awareness, ground the soul in life and existence. These ancient intuitions suggest we are fundamentally "more breath than thought and more meat than machine." Ultimately, research into consciousness has the potential to reposition us within the natural world, recognizing us as "living creatures with more in common with other animals than with the statistical abstractions of AI."
