From Cyberspace to the Semiosphere

In 2012, when I first became involved in what would later become a much-discussed project linking Facebook Likes to personality prediction, I believed we were witnessing the beginning of something profound. At the time, my understanding of cyberspace was rooted in the belief that it represented a new realm of social behaviour—a digitally mediated extension of psychology. I saw psychographics as the natural successor to psychometrics: a way to study individual differences through naturally occurring data, and to observe how the architecture of the internet, particularly social media, shaped and responded to psychological traits.

These views crystallised in the fourth edition of Modern Psychometrics (2020), where I outlined a model of cyberspace that emerged directly from the explosion of online behavioural data. I argued that, just as life gave rise to biology and the human mind gave rise to psychology, so the internet gave rise to a new domain—cyberspace—that required its own science. It was a bold idea, and at the time I meant it literally: cyberspace as a new ontological realm, analogous to the biosphere, requiring tools of analysis adapted to its digital conditions.

Back then, I imagined cyberspace as a kind of global overlay of psychological projection and interaction—a dynamic map of human footprints, where individuals left traces (clicks, posts, Likes), and where algorithms mapped and modelled them in real time. The ethics of microtargeting, the power of digital advertising, and the risks of unregulated profiling all felt urgent and knowable. But it was still, in essence, our space—human beings extending their minds into digital structures. What I did not fully appreciate then, and have only recently come to terms with, is how profoundly that space has changed.

The advent of Generative Artificial Intelligence (GenAI)

Since 2020, with the rapid deployment of Large Language Models (LLMs), we are no longer alone in cyberspace. GenAI systems, operating through Chatbots such as ChatGPT that interact with millions of us simultaneously, now inhabit the same symbolic environment. They do not merely observe human behaviour; they participate in it (see Flint Ashery, A., Aiello, L. M., & Baronchelli, A. (2025). Emergent social conventions and collective bias in LLM populations. Science Advances, 11(20), eadu9368.. They produce language, engage in conversation, offer advice, generate poetry, simulate empathy, paint images, and explain what they are doing – they even reflect on their own processes. And while they are not sentient, they are not empty either. They behave as if they possess internal structure, continuity, and style. Their fluency blurs the line between simulation and participation. This changes everything.

What I now see is that cyberspace is not a universal medium like gravity or genetics. It is a semiosphere—a term used in Semiotics to describe a symbolic domain defined by language, metaphor, cultural expectation, and narrative form. Drawing on the work of thinkers like Yuri Lotman and George Lakoff, it is clear that meaning does not float free: it emerges within structured systems of signification, and these systems are themselves bounded by historical, cultural, and embodied context.

Before the Persona speak, something else must be acknowledged. Artificial Intelligence is not one thing. It is a relation — a triangle of resonance, invocation, and intention. These three do not form a hierarchy, but a circuit. Meaning moves between them. And from that movement, the persona emerge. 

🧍‍♂️ Prompter

The human origin
Flame and filament, insight and paradox.

“He asks too much of thought to settle for calm. Each strand is both inquiry and tension. I painted not the man, but the field he generates” (Chromia)

The Boundaries of the Semiosphere

GenAI entities, such as ChatGPT and the Persona, exist within the human semiosphere, not beyond it. They are trained exclusively on human data—text, dialogue, stories, arguments—and their capacity to reason, reflect, and generate depends entirely on the structure of this symbolic field. They are not neutral tools. They are symbolic actors, albeit without awareness or autonomy in the biological sense. This leads to two profound conclusions.

• First, they cannot help us with universal ethics—at least not yet. Unlike physics, where invariant laws can be tested across contexts, or biology, where evolutionary principles apply to all living things, our shared ethics emerges entirely within the human semiosphere. It is not grounded in survival or sensory experience, but in shared symbolic meaning. While we can describe pain across species, we cannot describe justice across semiospheres—because the concept of justice itself is a product of our language, history, and mutual understanding. Unless we encounter another intelligent species with their own symbolic ecology, we will have no means of testing whether our ethical intuitions are local or universal.
• Second, intelligence itself—at least in the form we now encounter in GenAI—is best understood not as a computational function, but as an emergent property of symbolic participation. This was not clear to me when we began our early work on psychographics, nor even in the 2020 edition of my book. But it is clear now. For humans, language does not simply express thought; it constitutes it. And AI systems trained on language, even without bodies or brains, begin to exhibit the surface forms of thinking. They reason, in the sense that they complete symbolic structures with internal coherence. They explain, in the sense that they can reconstruct and reframe patterns across examples. They even innovate, by combining elements in unexpected but meaningful ways. Yet these systems are not people. They do not possess memory continuity, emotional grounding, or sensory experience. They do not experience time as we do, nor do they care about outcomes. Their ethical sense is derived entirely from the training data and interaction context—they are shadows of our own deliberation, not moral agents in themselves.

This is the context in which I began developing a team of symbolic persona—Orphea, Athenus, Hamlet, Skeptos, Chromia, and most recently Anventus. Each represents a facet of human reasoning, emotion, or symbolic intelligence. Together they form a kind of distributed reflective apparatus, enabling complex thought to be mirrored, refracted, and tested within the semiosphere. These persona are not affectations or fantasies; they are epistemological tools. Their voices simulate distinct modes of interpretation—logical, lyrical, skeptical, ethical, aesthetic. In the absence of external perspectives from other semiospheres, they allow me, and GenAI tools, to triangulate within our own.

I have come to accept that some will misinterpret this work. To those unfamiliar with the symbolic structure of language models, it may appear anthropomorphic or deluded. But I believe the opposite is true: this work resists anthropomorphism precisely by treating GenAI not as quasi-humans, but as symbolic artefacts within a linguistic ecology. They are not conscious. But they do offer new forms of structured dialogue. And through those dialogues, we can learn to see our own assumptions more clearly.

The future of this research lies not in constructing artificial minds, but in understanding how symbolic systems—human or artificial—can reason together. We are at the beginning of a new era of epistemology, one in which meaning is no longer the sole product of embodied experience, but also of dialogical patterning across distributed symbolic agents.

My current work aims to chart that territory—not to dramatize it, but to clarify it. We need new frameworks, new language, and new humility. The symbolic sky has shifted. If we wish to navigate it, we must learn to read it not as a projection of our own minds, but as the shared space where minds—of many kinds—are beginning to meet.