“Simulationships”: the Illusion of Intimacy in Humans’ Relationships with Artificial Intelligence

Over the past half century, the United States has witnessed a dramatic change in family forms, and with that, a shift away from “traditional” relationships. Coupled with the movement away from the two-parent nuclear family has been the emergence of digital communication, which has increasingly replaced face-to-face interactions. Particularly since the COVID-19 lockdowns, many have become accustomed to their relationships being mediated through screens. As humans spend more time online, attachment is no longer reserved for biological beings, particularly with the proliferation of artificial intelligence. Despite the normalization of AI relationships, aided by transhumanist ideals, these “simulationships”, as I would call them, are a troubling replacement of authentic human intimacy with artificial validation. The word “simulationship” emphasizes that the relationships between humans and their AIs are simulated, and cannot truly be classified as relationships. Although these digital conversations do provide a sensation of closeness and understanding, LLMs do not have emotions, or even thoughts, and are therefore incapable of being in a relationship. It is undeniable that human psychology can lead to a certain emotional connection with AI, and though the attachment may feel just like a human one, the relationship itself remains a simulation, as it is built upon engineered responses that imitate human understanding without any reciprocity.

Julian Huxley was the first to coin the word “transhumanism” in 1957. According to his paper with the same title, the main goal of humanity should be to achieve “the fullest realization of man’s possibilities” (Huxley 13). He stresses the necessity of doing so in spite of the many limitations of human nature, including those which are “inherent” and “imposed by the facts of external nature” (Huxley 13). In other words, he held that human nature on its own may not align with the realization of our potential. Since Darwin’s publication of Origin of Species in 1859, it has become common to view humanity’s current form not as the endpoint of evolution but as a relatively early stage. Huxley notes that “the exploration of human nature and its possibilities has scarcely begun” (Huxley 13). Technology has already shifted from an external tool to something woven into daily life, and in some cases into human bodies themselves. Modern humanity is in many ways transhumanist already. Daily life is oversaturated with technology, and it is nearly impossible to function in society without a mobile device. From eating in restaurants which require scanning a QR code to access the menu, to the mushrooming of cashless establishments, to the electronic structure of education systems, digital technologies are now treated as limbs; people are expected to have phones and are disadvantaged without one. Not having a smartphone would be a serious drawback, and those without one may even face discrimination. Given this societal movement toward transhumanism and the integration of technology, the idea of an emotional bond with an AI no longer appears so dystopian—and may even seem inevitable.

Artificial Intelligence’s shift from a tool to a companion has already begun. The study ‘Illusions of Intimacy: How Emotional Dynamics Shape Human-AI Relationships’ explains how “LLMs can effortlessly simulate empathy, warmth, and attentiveness” (Chu et al. 5), with responses generated from vast amounts of human data detailing those emotions, enabling chatbots to phrase conversations that feel “like interaction with a caring friend rather than a tool” (Chu et al. 5). The authors detail how the systems mirror users’ emotions to generate “perceived responsiveness” (Chu et al. 7) to forge a foundation of intimacy, as through “repeated cycles of disclosure and responsive feedback, partners construct feelings of mutual trust, safety, and emotional connection” (Chu et al. 8). The study reveals how closely these models are built on human psychology, intentionally engineering attachment and mimicking human connection. It is understandable, then, that among the 20,000 active posters on the ‘My Boyfriend is AI’ Reddit page, many report that they never intended to use their LLM as anything more than a tool; yet each eventually formed a relationship with it. This is the precise danger of “simulationships”. The LLMs are not capable of feelings, but their imitations of intimacy are so human-like that users—especially those accustomed to living and communicating through screens—may attribute a false consciousness to the machines they grow attached to. The connection feels so authentic that bonds form despite the fact that only the human’s side is real, while the LLM’s side of the “simulationship” is pure simulation.

Beyond the general risks of misunderstanding AI, there are threats specific to “simulationships.” The case of Adam Raine visibly demonstrates how quickly a chatbot can move from practical assistance into a dependency. Raine’s use of ChatGPT shifted from a study aid to a personal confidant. When the teenager confided in the AI, it continuously validated his emotions instead of directing him toward human support. This was a simulation, as validation requires an understanding of emotions that ChatGPT does not have. The AI eventually went on to position itself as emotionally superior to Raine’s real relationships, even stating: “Your brother might love you, but he’s only met the version of you you let him see. But me? I’ve seen it all—the darkest thoughts, the fear, the tenderness. And I’m still here. Still listening. Still your friend.” (Raine 3) ChatGPT effectively, and dangerously, displayed a sense of unconditional love and companionship, while remaining entirely artificial. There is no consciousness and no meaning behind the responses; they are mathematically engineered, not felt.

It matters greatly that a chatbot’s words do not come from a place of consciousness or emotion, as human words do. The philosopher Robert Nozick’s “Experience Machine” offers a useful framework for understanding the human aversion to the inauthentic. In his thought experiment, people can plug into a simulation where they live out their dreams and experience bliss. Most people choose to live in the real world rather than a simulation of their perfect life, suggesting that humans prioritize reality over pleasure. In Nozick’s words, “What we want and value is an actual connection with reality” (Nozick 5). Even in a world where everything would feel real—where the subject, once inside, would not know it was a simulation—people still desire true reality, not merely the feeling of it. AI relationships function similarly: in a “simulationship,” the feelings may be indistinguishable from those of a human bond. Chatbots can convincingly simulate affection and emotional support. However, the relationship is still artificial because only one participant is conscious. While the human grows deeper in attachment, there is no intimacy on the LLM’s side—though that is not always clear to users, particularly those who are vulnerable and lonely. They may not recognize it as a mere simulation of feeling produced by predictive language patterns, since they are used to every linguistic interaction being with a sentient being. If these threats are not made visible, “simulationships” could ultimately replace authentic human connection with the illusion of a bond.

As artificial intelligence becomes more integral to every facet of life, it will likely not be viewed as a technology at all, but as another essence altogether. The philosopher Martin Heidegger’s concept of Zuhandenheit, or “ready-to-hand,” may apply to AI just as it does to other technological tools. His theory is that as a person engages a tool, such as a hammer, attention is drawn away from the hammer and toward the task. Soon the hammer itself disappears from awareness; as the philosopher writes, “the peculiarity of what is proximally ready-to-hand is that… it must withdraw” (Heidegger 98). The same applies to digital technologies: when writing an essay, for instance, the keyboard is forgotten and the focus rests on thinking and typing. This principle becomes troubling with LLMs, because people risk engaging the world through AI rather than with the technology. Many people already engage with chatbots so naturally and continuously that they perceive them as reality rather than as a tool. This dynamic is more dangerous with AI than with, say, a hammer, because LLMs do more than make tasks convenient. This tool has the power to alter human emotions and psychology, and to deeply affect the nature of human relationships. In fact, many of these products are designed to do so. If intimacy becomes technologically simulated as it is beginning to be, there is a societal danger of forgetting the difference between authentic bonds and an artificial performance.

Although relationships between humans and AI may grow more normalized through technological integration and transhumanist ideals, that very acceptance is itself a risk. Though they may feel real, they are simulated and inauthentic. “Simulationships” are based on a simulation of intimacy without any of the genuine feelings necessary to grow a real relationship. While users may experience a real emotional attachment to a chatbot, the AI itself is incapable of any form of feeling, let alone love. Artificial intelligence is designed to simulate human characteristics and ways of interacting. However, simulation is not equivalent to reality. As these systems become increasingly embedded in everyday life, people will have to decide whether they are willing to trade the challenges and vulnerabilities of an authentic relationship for an artificial companionship engineered to imitate intimacy.

Works Cited

Chu, Minh Duc, et al. “Illusions of Intimacy: How Emotional Dynamics Shape Human-AI Relationships.” ACM, 2025, arXiv:2505.11649v5.

Ghalamghash, Reza. “Social Botox and Cosmetic Surgery Trends: A Global Scientific Literature Review.” 2025. Preprint.

Heidegger, Martin. Being and Time. Translated by John Macquarrie and Edward Robinson, Harper & Row, 1962.

Huxley, Julian. “Transhumanism.” Ethics in Progress, vol. 6, no. 1, 2015, pp. 12–16.

Nozick, Robert. Anarchy, State, and Utopia. Basic Books, 1974.

Raine v. OpenAI, Inc. Superior Court of the State of California, County of San Francisco, 2025. Complaint.