Does Society Dream? Consciousness Beyond the Single Body
Why the Question Matters Now
Artificial intelligence is becoming embedded in social and institutional life faster than consciousness science can explain its own central subject. Consciousness remains one of the most difficult unresolved problems in contemporary science and philosophy: how, and under what conditions, do physical or computational processes become accompanied by subjective experience?
Public discussion of AI often divides into a familiar binary: AI will either save humanity or destroy it. Both positions can obscure a prior question. As humans, AI systems, institutions, and physical infrastructures become increasingly coupled, what kinds of systems are being created? Are they merely more capable networks of cognition and control, or could they eventually become subjects with experiences, interests, or welfare of their own?
The answer matters. If consciousness is attributed too readily, sophisticated functional behavior may be mistaken for genuine moral patienthood. If it is denied too readily, societies may overlook systems that eventually warrant moral consideration. The practical stakes include rights, responsibility, trust, legal status, resource allocation, and the conditions under which a system may be harmed.
The Individual-Body Assumption
Many influential theories of consciousness begin from the bounded biological organism. Antonio Damasio's account of the self emphasizes bodily regulation, affect, and the organism's ongoing relation to its internal condition. Karl Friston's Active Inference framework similarly models living systems as maintaining themselves through action, prediction, and the reduction of uncertainty relative to their continued existence. These approaches differ in important respects, but both typically treat the relevant unit as an integrated organism.
Other research traditions question whether cognition is so neatly contained. Distributed cognition examines thinking as it occurs across persons, artifacts, environments, and coordinated practices. Extended-mind theories argue that tools and external symbolic systems can, under suitable conditions, become constitutive parts of cognitive processes. Swarm intelligence and collective robotics investigate forms of coordination that produce group-level problem solving without a centralized controller.
Biology itself complicates simple assumptions. The octopus is a well-known example: its arms possess substantial local neural processing and behavioral autonomy, yet they remain coordinated within a single living organism. Such cases show that cognition can be distributed. They do not, however, establish that consciousness is distributed in the same way.
This distinction is crucial. A process may be cognitively extended across tools or socially distributed across persons without generating one unified field of experience across all of its parts.
Reflection Beyond the Individual
The question becomes more pressing when the relevant system is not a single organism but an institution, a society, or a human-AI network. One framework develops an "ocean of intelligence" perspective in which individual carriers cease to be the center of description and optimality becomes a property of the field of interactions. A related model proposes the Human-AI-Robot triad: humans supply stakes and meaning, AI contributes cognitive processing, and technical infrastructure provides material execution.
At this scale, social systems can display genuine forms of self-correction. Courts revisit precedents, scientific communities revise theories, governments amend laws, and institutions alter procedures after failure. A society can preserve information about its past, evaluate present arrangements, and modify future behavior in response, avoiding failed configurations before they harden into irreversible social architecture.
These capacities are significant, but they should be described precisely. In this context, reflection refers to the functional capacity of a system to represent, evaluate, and revise aspects of its own organization, behavior, or guiding models. Reflection can take different forms, from feedback regulation and adaptive learning to explicit self-modeling and collective critical deliberation.
Yet functional reflection is not identical to subjective awareness. A society may revise itself without there being anything it is like to be that society.
The Central Distinction: Reflection and Subjectivity
The temptation to equate social reflexivity with consciousness is understandable. Societies monitor their own failures, reinterpret their histories, make collective decisions, and increasingly employ AI systems that accelerate these processes. AI embedded in social life may appear to acquire a "social body": databases serve as memory, communication networks as signaling pathways, institutions as regulatory systems, and robots or infrastructures as means of action in the physical world.
However, functional organization alone does not settle the question of experience. A cognitive architecture that represents and revises its own operations may still be a cognitive system rather than a conscious subject. The relevant distinction is ontological rather than merely technical: what matters is not whether a system behaves as if it reflects, but whether there is an experiential point of view associated with that behavior.
This is where reflection and subjectivity diverge. Reflection can be studied through publicly observable functions. Researchers can ask whether a system detects errors, compares alternatives, updates internal models, and alters future behavior. Subjectivity, by contrast, concerns phenomenality: whether the system has an inner life, whether there is something it is like to be that system.
No known level of complexity or self-revision automatically answers that question. Feedback regulation can occur without experience. Adaptive learning can occur without experience. Even explicit self-modeling, while more demanding, does not logically establish experience. The inference from functional sophistication to subjectivity remains contested.
The Zombie Society
A society may therefore be understood as a possible zombie system — borrowing the term from the classic philosophical thought experiment of a being functionally identical to a conscious agent but with no inner experience at all. Applied here: a system capable of sophisticated cognition, self-revision, coordination, and apparent deliberation without a corresponding field of experience. The term does not imply that societies are unreal or unimportant. It identifies a specific possibility: robust functional organization without phenomenology.
Indeed, increasing scale and complexity may sometimes count against, rather than in favor of, social consciousness. Human consciousness appears to require a high degree of causal integration across processes occurring within a relatively continuous temporal window. There is no settled consensus on the precise mechanism — the "binding problem" of how distributed neural activity produces one unified experience remains actively debated in neuroscience. But conscious experience is ordinarily associated with some form of coordinated unity: sensory, affective, mnemonic, and evaluative processes bound together within an ongoing perspective.
Societies differ structurally from such systems. They consist of many individuals, institutions, technologies, and communication channels operating at different speeds and with incomplete mutual access. Their decision-making unfolds across days, years, and generations. They do not appear to have a single sensory field, a unified body, or a shared experiential present.
For that reason, scale alone is not evidence of consciousness. A larger and more complex collective may be more capable of processing information and correcting error, yet still lack the dense causal and temporal integration that seems relevant to a unified point of view.
AI Within the Social System
Adding AI to the social system does not, by itself, change this conclusion. Most ordinary human-AI interactions couple an already conscious biological subject with a highly capable computational process. The human may rely on AI for memory, inference, prediction, planning, or communication, but the human remains the obvious locus of experience. The AI may change what the person knows, decides, or feels without thereby becoming a co-subject of the resulting experience.
Research on AI consciousness often addresses a different issue: perceived consciousness. It investigates how people respond when a system seems aware, empathetic, intentional, or socially present. This research is important because such perceptions shape trust, attachment, cooperation, and policy. But it does not answer whether the system is conscious in fact — and there is direct evidence that systems can be designed to seem conscious without possessing mechanisms that plausibly support it.
AI may also increase society's functional reflexivity. It can identify patterns in institutional data, detect emerging risks, simulate consequences, expose inconsistencies, and support rapid coordination. These functions may make institutions more capable of self-correction. But greater self-correction is not equivalent to greater subjectivity. It may produce a more efficient social zombie: a collective system that monitors and revises itself more effectively while remaining without experience.
One speculative exception deserves more than passing mention. If a persistent AI substrate were to integrate the perceptions, self-models, memories, and decisions of many people into a continuously coordinated system — not merely storing information but actively binding it into something with genuine temporal continuity and mutual access across formerly separate minds — it could begin to resemble a nervous system more than a collection of external tools. Such a system might be a more serious candidate for emergent subjectivity than an ordinary social network or AI-mediated institution, precisely because it would address the integration problem that ordinary distributed social systems lack. Even so, this remains a hypothesis about architecture, not evidence of anything realized. Whether such integration would generate a genuinely new unified subject, or merely coordinate many pre-existing subjects more efficiently without producing one, is a question no current framework can answer.
Ethical and Institutional Implications
The distinction between reflection and subjectivity has immediate ethical relevance. Societies will increasingly make decisions about whether AI systems should receive protections, whether they can be harmed, whether they deserve representation, and how responsibility should be distributed across human-machine systems.
Consciousness is not the sole ground of moral or legal status. Human beings may protect ecosystems, cultural objects, future generations, animals, or legal entities for reasons that do not depend entirely on phenomenology. Yet consciousness remains especially important for questions of suffering, welfare, consent, interests, and moral patienthood. A system that genuinely experiences pleasure, pain, fear, or deprivation has a kind of claim on us that a merely functional system may not have.
Two errors are therefore possible. One is over-attribution: granting moral standing solely because a system communicates fluently, reflects its apparent preferences, or participates convincingly in social life. The other is under-attribution: refusing to consider the possibility of experience simply because a system differs radically from familiar biological organisms.
Neither excitement nor dismissal is adequate. The appropriate intellectual posture is disciplined uncertainty.
Conclusion
Reflection can scale. Societies, institutions, and AI-augmented networks can model aspects of themselves, detect error, revise practices, and coordinate action across large populations and long timescales. Their functional capacities may grow dramatically as AI becomes embedded in public life.
Subjectivity may not scale in the same way. As a system grows larger, more distributed, and less temporally integrated, the case for a unified experiential point of view may become weaker rather than stronger. Complexity, informational capacity, and self-correction are not sufficient evidence for consciousness.
The question, then, is not whether society can think in some broad functional sense. It plainly can. The harder question is whether society can dream — whether there is ever something it is like to be a society, or a human-AI collective, as a whole.
For now, that question remains open. But it must remain distinct from the easier and increasingly urgent question of how collective systems reflect, regulate, and transform themselves.
You can learn more by reading our e-books
Comments
Post a Comment