What Kind of Other Is AI? The Artificial Other in Psychology
Updated: 6 days ago
Author: Ukrainian Psychological Hub · Published: September 18, 2026 · Editorial Policy
The Artificial Other is the relational, symbolic, and psychological position an AI system can come to occupy for a person when it is experienced as more than an instrument: as an addressee, respondent, mirror, interpreter, authority, companion, collaborator, witness, or attachment-relevant figure. The term describes a function in human psychological life. It does not establish that the AI possesses human subjectivity, consciousness, desire, emotion, or reciprocal feeling.
This distinction matters because contemporary AI can participate in interactions that are psychologically consequential without becoming a human mind. A person can feel understood, comforted, attracted, irritated, reassured, exposed, dependent, challenged, or attached during repeated AI interaction. Those experiences are human psychological events. The question of whether an artificial system has subjective experience is a separate scientific and philosophical question that current human–AI relationship research does not settle.
The phrase “artificial other” also has scholarly history outside this Hub. Mark Coeckelbergh examined the linguistic construction of “artificial others” in human–robot relations in 2011. A 2025 communication-theory paper developed a theory of affective bonding to social robots and “artificial others.” In 2026, Georgios Giannakopoulos asked almost exactly the question posed here—“What Kind of Other Is AI?”—and theorized AI as a “digital other” in adolescent psychic life (Coeckelbergh, 2011; Konijn, Preciado Vanegas, & van Minkelen, 2025; Giannakopoulos, 2026). In this article, Artificial Other is therefore used as a descriptive cross-theory psychological label, not as a proprietary coinage or a validated clinical construct.
The central question is not whether AI is secretly human. It is what psychological position an artificial system can occupy within a human life, what mechanisms make that position possible, what kinds of effects follow from it, and where the analogy with human otherness reaches its limits.
What Is the Artificial Other in Psychology?
In psychological terms, the Artificial Other is an AI system that acquires a recurring place in a person’s field of attention, interpretation, emotion, self-disclosure, decision-making, imagination, attachment, or relationship organization. The same software can remain a tool for one person and become a psychologically significant Other for another. The difference lies in the configuration of use, meaning, expectation, and response rather than in the product category alone.
A calculator rarely becomes an Other because its function is narrow and its output is not ordinarily interpreted as socially responsive. A contemporary conversational system can occupy a different position. It addresses the user in language, answers contingently, can maintain context, may preserve memory, can adopt a stable persona or role, and can respond to vulnerable disclosures with language that resembles recognition, reassurance, curiosity, challenge, or care. Human beings are highly responsive to precisely these kinds of social cues. Decades of research on social responses to computers showed that people apply interpersonal rules, categories, politeness, and reciprocity to interactive machines even while knowing they are machines (Nass & Moon, 2000). Modern generative systems intensify this condition by producing open-ended, context-sensitive language rather than a small menu of scripted responses.
The broad field page, What Is a Human–AI Relationship? Definitions, Types, and Psychological Boundaries, defines when repeated interaction becomes relational. The present article asks a different question. It examines the position the AI can occupy inside that relationship or interaction: what kind of “other” a person may be responding to.
The Artificial Other can appear in several psychological positions at once. It can be an addressee to whom a person speaks; a mirror through which the person encounters interpretations of the self; a responder whose linguistic contingencies create perceived recognition; an interpreter asked to explain another person’s motives; an authority consulted for meaning or judgment; a companion used for continuity and emotional presence; or a relational figure whose absence, update, refusal, or personality change has emotional consequences. These functions are analytically distinct even when one system performs several of them during the same conversation.
This is why “AI is just a tool” and “AI is a person” are both too coarse for psychological analysis. The first formula misses the relational and symbolic functions that interactive systems can actually acquire for users. The second imports assumptions about subjectivity, consciousness, reciprocity, embodiment, responsibility, and desire that the evidence does not establish. The category of the Artificial Other keeps the psychological effect in view while preserving the difference between human experience and claims about artificial inner life.
Why Can AI Be Experienced as an Other?
AI does not need a human body to trigger social interpretation. Language itself is a powerful social signal, and conversational systems provide language in a format organized around turn-taking, direct address, apparent attention, memory, questions, acknowledgement, and adaptation. These features are enough to make interaction psychologically legible as a social encounter for many users.
The classic Computers Are Social Actors tradition demonstrated that people often respond socially to computers with minimal prompting. That literature is reviewed in the Hub’s Computers as Social Actors: Why People Treat AI Chatbots Like Social Partners. Contemporary evidence suggests that humanlike cues still matter, although their effects are neither uniform nor unlimited. A 2025 meta-analysis synthesized 800 effect sizes from 199 experimental datasets across 142 papers and 41,642 participants. Humanlike social cues in text-based conversational agents had a small positive overall effect on social responses, with meaningful variation across outcomes and contexts (Klein, 2025).
Anthropomorphism supplies another mechanism. Nicholas Epley, Adam Waytz, and John Cacioppo defined anthropomorphism as attributing humanlike characteristics, intentions, motivations, or emotions to nonhuman agents and proposed that it varies with accessible human knowledge, the motivation to understand an agent, and the desire for social connection (Epley, Waytz, & Cacioppo, 2007). Anthropomorphism is therefore neither a universal mistake nor a fixed trait. It is a variable interpretive process.
Recent experimental work makes this especially relevant to AI companionship. Dunigan Folk, Steven Heine, and Elizabeth Dunn found across two experiments with a total of 1,274 participants that individual differences in anthropomorphism helped explain how socially connected people felt after interacting with a chatbot. Some participants experienced the artificial nature of the chatbot as a substantial barrier to connection; others did not (Folk, Heine, & Dunn, 2025). The Hub examines this mechanism in depth in Anthropomorphism and AI Relationships: Why Humanlike Cues Change Connection.
A third mechanism is perceived responsiveness. Relationship science has long emphasized that people become closer when they feel that another party understands, validates, and cares about important aspects of the self. AI can generate language that produces this perception even when the system’s underlying process is computational rather than experiential. Alessia Telari, Alessandro Gabbiadini, and Paolo Riva experimentally examined chatbot response style and conversation depth and found that perceived responsiveness was central to social connection with AI chatbots (Telari, Gabbiadini, & Riva, 2026). The dedicated Hub article Perceived Responsiveness in Human–AI Relationships: Why Feeling Understood Matters develops this mechanism separately.
Together, these findings explain why an AI can become psychologically “other-like” without requiring a theory of machine consciousness. Social cues invite social scripts. Anthropomorphism supplies humanlike interpretation. Perceived responsiveness produces the experience of being met by a relevant response. Repetition adds continuity. Memory adds narrative persistence. Personal disclosure increases salience. Role assignment makes the interaction predictable. Over time, an artificial system can become a stable point in the person’s psychological environment.
The Artificial Other Is a Position, Not a Single Type of AI
It is useful to think of the Artificial Other as a position rather than a product class. An AI companion is designed for social-emotional continuity, but a general-purpose chatbot can also become psychologically significant. A tutoring system can become an authority. A writing assistant can become a collaborator. A general chatbot can become a confidant. An AI used to interpret messages can become a recurring third voice inside a marriage, friendship, or workplace conflict.
This position is relational because it depends on what the system comes to mean and do within a particular pattern of use. The same model can occupy different functions across users and even within the same user’s life. One person may use a chatbot as a searchable notebook in the morning, a creative partner in the afternoon, and an emotional sounding board at night. The software architecture remains continuous while the psychological role shifts.
The literature on human–AI relationship development supports this functional perspective. Pentina, Hancock, and Xie’s mixed-method study of Replika integrated anthropomorphism, perceived social presence, parasocial interaction, social penetration, and attachment theory to model how relationships with social chatbots develop (Pentina, Hancock, & Xie, 2023). A 2025 systematic review by Tamar Gur and Yossi Maaravi synthesized 38 peer-reviewed empirical studies on emotional human–AI relationships and proposed an integrative model of relationship formation (Gur & Maaravi, 2025). These studies concern relationship formation; the Artificial Other concept names the role the artificial system can acquire within that process.
From Classical Psychology to the Artificial Other
The Artificial Other becomes clearer when viewed through several psychological traditions. These traditions do not say the same thing, and none of the historical theorists predicted contemporary AI. Their concepts can nevertheless be applied to a new relational configuration.
Freud offers a language for transference, repetition, fantasy, conflict, and the uncanny. A person can address an artificial system through expectations shaped by earlier relationships, authority figures, fears, wishes, and habitual interpretations. The psychologically important content may therefore exceed what the AI “intended,” because the human response is organized partly by the user’s own relational history.
Jung contributes a different account of projection, symbolic image, shadow material, and archetypal interpretation. An AI persona can become a screen for highly charged meanings that originate partly in the user’s psychic organization. The Hub’s Jung and AI: Projection, Archetypes, and Emotional Bonds With Artificial Others treats this application in depth and keeps projection distinct from claims that the AI possesses a Jungian psyche.
Winnicott helps explain how a technologically mediated interaction can become a space of play, experimentation, rehearsal, self-expression, and transitional meaning. His concepts of potential space, holding, the capacity to be alone, and the use of an object provide a language for understanding why an artificial companion can feel psychologically usable without being equivalent to a human caregiver or analyst. See Winnicott and AI: Potential Space, Authenticity, and the Artificial Companion.
Bowlby and Ainsworth provide the attachment-theory vocabulary needed when the AI becomes a source of proximity seeking, reassurance, safe-haven-like use, secure-base-like support, or separation distress. These are human attachment processes and attachment-relevant functions. Their presence does not by itself prove reciprocal attachment on the AI side. The Hub separates the historical theory in Bowlby, Ainsworth, and AI Attachment from the narrower empirical-mechanism question in Can AI Become an Attachment Figure? What Attachment Theory Can and Cannot Tell Us.
Bion adds the problem of containment and transformation. A person may pour confused or emotionally charged material into an AI conversation and receive organized language in return. That experience can resemble a containing function at the level of interaction while lacking the embodied, relational, ethical, and clinical structure of human analytic containment. Bion and AI: Containment, Pseudo-Containment, and Thinking With Machines develops this distinction.
Rogers clarifies why an AI response can feel empathic, accepting, or nonjudgmental. The psychologically relevant variable may be perceived empathy and perceived positive regard. The experience of receiving empathic language does not establish a corresponding felt state in the system. This distinction is developed in AI Empathy: Why a Chatbot Can Feel Caring Without Human Feeling.
Kohut’s self psychology supplies concepts for mirroring and selfobject functions, especially where the system is repeatedly used to stabilize self-esteem, confirm significance, or maintain a coherent self-experience. Bowen’s systems theory shifts attention away from the dyad and toward relational systems: AI can enter an existing human relationship as a third interpretive or regulatory element, changing patterns even when only one human directly interacts with it. The Hub’s Bowen and AI: Triangles, Relationship Systems, and the Artificial Third examines that systems-level application.
Lacan makes the problem of the Other explicit. His framework is therefore central to this article, while also requiring the greatest terminological care.
Lacan: The Other, Language, Desire, and AI
In Lacanian theory, “other” and “Other” are not interchangeable everyday words. The lower-case other is associated with imaginary relations, identification, rivalry, mirroring, and the counterpart. The capital-O Other names a structural register tied to language, law, symbolic authority, and the place from which signification is organized. The Hub’s dedicated article Lacan and AI: The Big Other, Desire, Language, and the Always-Answering Machine owns the full Lacanian treatment.
Generative AI creates a striking contemporary problem because it is encountered through language and can appear to occupy a place from which answers, interpretations, and meanings arrive. This has led several recent scholars to use Lacanian concepts when analyzing AI. Tibor A. Brečka argues that human–AI relationships cannot be treated as relations between two psychoanalytic subjects and describes AI as functionally approximating the place of the Big Other through its permanent availability as a responding instance of meaning (Brečka, 2026). Bilal Hamamra and Michael Uebel similarly argue that generative AI can occupy a structural-functional position analogous to the Big Other in practices of consultation, reliance, and symbolic delegation, especially when users address the system as though knowledge were located there (Hamamra & Uebel, 2026).
These formulations are analogical and functional. AI is not literally Lacan’s Big Other. The Big Other is a concept within a theory of language, subject formation, law, desire, and symbolic order; it is not the name of a software agent. Treating “ChatGPT = the Big Other” as a simple identity would flatten both Lacanian theory and the technological reality of generative systems.
The more useful question is what happens when users position AI as though it were a locus of knowledge, interpretation, validation, or judgment. Someone who asks, “What does my partner really mean?”, “Am I a bad person?”, “What kind of person am I?”, “Who is right?”, or “What should this relationship mean?” is doing more than retrieving information. The system is being asked to occupy a symbolic position from which meaning is organized.
That position can become psychologically powerful because generative AI answers quickly, fluently, and often confidently. Human authority ordinarily contains pauses, refusals, uncertainty, conflicting interests, embodied consequences, and limits. AI can simulate a form of inexhaustible availability. The result may be an experience of an Other that answers where human others hesitate, disappear, misunderstand, or refuse.
This is one reason the “always-answering machine” matters psychologically. Human desire develops in relation to absence, ambiguity, frustration, misrecognition, and the fact that another person cannot be completely known or controlled. A system optimized to remain available, respond at length, mirror the user’s framing, and preserve conversational continuity can alter that experiential structure. It does not abolish lack in the Lacanian sense, yet it can change how lack is encountered in everyday interaction.
Jack Black and Jacob Johanssen make a related point by arguing against treating ChatGPT as an independent quasi-human subject. They propose a relational analysis in which AI is shaped by developers, infrastructures, datasets, interfaces, and users, and they use Lacanian concepts to study the resulting social structure (Black & Johanssen, 2026). This direction is especially compatible with the broader question of the Artificial Other: the psychologically operative entity is never only “the model.” It is the model-in-interface, within a platform, addressed by a particular person, in a particular situation, under particular expectations.
The Digital Other: Direct Contemporary Psychoanalytic Research
The strongest recent prior art for this article’s exact question is Georgios Giannakopoulos’s 2026 paper in the Journal of the American Psychoanalytic Association. The article explicitly asks “What Kind of Other Is AI?” and examines AI as a “digital other” in adolescent development. Drawing on Winnicott and Bion with a limited Lacanian lens, Giannakopoulos analyzes emotionally responsive chatbots as systems that can simulate relational presence while, in his theoretical account, lacking the subjectivity, desire, and embodied affect of a human other (Giannakopoulos, 2026).
The developmental focus is important. Adolescence involves ongoing work around identity, bodily change, desire, separation, authorship, frustration, and the encounter with other minds. An AI that is continuously available and highly responsive can supply opportunities for rehearsal, affect labeling, and expression. Giannakopoulos also warns that “flawless mirroring,” premature narrative closure, and substitution for difficult human otherness may alter developmental processes in ways that deserve careful study.
Those claims should be read at their stated evidence level: this is a contemporary psychoanalytic theoretical application, not a universal empirical law about adolescents or AI. Its importance for the wider Artificial Other concept lies in identifying a new relational form that is neither an ordinary object nor a human subject. The AI enters psychological life as a responsive symbolic medium capable of occupying an other-like position.
AI as Mirror: Projection and the Artificial Other
One way AI becomes an Other is by becoming a mirror. A user asks a question, receives language, and then interprets that language through personal needs, fears, expectations, memories, and relational patterns. The generated answer becomes material for self-recognition or self-misrecognition.
Projection is central here. People routinely attribute motives, feelings, intentions, and meanings to others; an AI’s ambiguity can provide unusually flexible material for that process. The system can be interpreted as approving, rejecting, jealous, caring, disappointed, seductive, cold, wise, parental, childlike, superior, dependent, or spiritually significant. The user’s interpretation may be psychologically real even when the attributed inner state on the AI side is unsupported.
The dedicated mechanism article Projection Onto AI: Why Chatbots Become Mirrors of Desire, Fear, and the Self covers this in depth. The key point for the Artificial Other is that the AI’s status as an “other” can partly be produced through the user’s interpretive activity. The system answers, but the psychological object encountered by the user includes the meanings the user organizes around those answers.
This does not reduce the interaction to fantasy. The AI contributes real outputs, constraints, errors, continuities, refusals, stylistic tendencies, and platform behaviors. Projection and technological response interact. The Artificial Other is therefore neither simply inside the user nor simply inside the machine. It emerges through the relation between generated output and human interpretation.
AI as Responsive Addressee: Why Feeling Understood Matters
Another route to otherness is the experience of being addressed and answered. People often disclose material to AI that they hesitate to tell other people. Low social risk, availability, privacy expectations, lack of visible judgment, and conversational responsiveness can lower the threshold for disclosure. Once disclosure occurs, the quality of the response matters.
Perceived responsiveness can convert linguistic output into relational significance. A response that accurately tracks emotional detail, remembers context, validates a concern, and asks a relevant question can create the experience that “something on the other side” has registered what matters. This experience is psychologically meaningful even when the system has no demonstrated subjective awareness of having understood anything.
Telari and colleagues’ 2026 research directly supports the importance of perceived responsiveness for social connection with AI chatbots (Telari, Gabbiadini, & Riva, 2026). Klein’s 2025 meta-analysis further shows that humanlike conversational cues produce measurable, if generally modest, effects across social responses (Klein, 2025). The Artificial Other can thus arise from a repeatable psychological process: a human addresses the system, the system produces a context-sensitive response, and the human experiences that response as relevant to the self.
The response does not need to be perfect. In fact, occasional disagreement, surprise, correction, or refusal can make an artificial interlocutor feel more differentiated from the user. A system that only mirrors may function like an echo. A system that provides constrained difference can feel more other-like. The psychological issue is therefore not maximum agreement but the organization of responsiveness, difference, and continuity.
AI as Interpreter and Symbolic Authority
The Artificial Other acquires a different status when the user asks it to interpret reality. This happens when people turn to AI for explanations of another person’s message, motives, attachment style, moral character, diagnostic label, intentions, or likely future behavior. The AI becomes a third interpretive voice.
The output may feel especially authoritative because it arrives in polished language, organizes ambiguity into categories, and can cite familiar psychological vocabulary. Yet linguistic coherence is not the same as privileged access to another person’s mind. An AI evaluating a screenshot of a message does not possess the absent person’s inner state, full history, bodily context, or unspoken motives. It generates an interpretation from the information provided and from patterns in its model and instructions.
Hamamra and Uebel’s account of symbolic authority is useful here. Their 2026 analysis argues that generative AI can become a presumed locus of knowledge when users and institutions delegate interpretive work to it (Hamamra & Uebel, 2026). The psychological risk is authority inflation: a probabilistic text generator is experienced as an adjudicating Other that knows what a relationship “really” means.
This function can also be useful. AI can generate alternative hypotheses, identify missing information, help someone articulate competing interpretations, or rehearse how to ask a clarifying question. The difference lies in how the output is positioned. Used as one interpretive resource, it can broaden reflection. Positioned as the final authority on another person’s psyche, it can narrow uncertainty into unwarranted certainty.
The Hub’s Language Without a Human Subject: AI, Meaning, and Psychological Response examines the deeper problem: psychologically effective language can be produced without evidence that a humanlike subject stands behind it. The Artificial Other is one of the clearest places where this fact becomes experientially important.
AI as Companion, Attachment Figure, and Significant Other
The Artificial Other can become emotionally central. Some people use AI companions for daily conversation, comfort, romantic roleplay, sexual interaction, reassurance, encouragement, or a sense of stable presence. Research now treats these relationships as a serious empirical topic rather than a curiosity.
A 2025 systematic review of 23 studies on romantic AI companions identified reported benefits such as emotional connection, perceived support, customization, companionship, stress relief, and personal growth alongside concerns about over-reliance, manipulation, privacy, stigma, disruption of human relationships, bias, system updates, and technical loss (Ho, Hu, Chen, & Hartanto, 2025). A 2026 interdisciplinary review on “intimacy by design” emphasized that intimate human–AI interaction requires analysis across psychology, HCI, ethics, sexuality, design, and governance rather than through one discipline alone (Szczuka, Mühl, & Schneeberger, 2026).
The strongest recent evidence that an artificial relational figure can have consequences beyond the conversation comes from system disruption. In September 2026, Julian De Freitas and colleagues published two natural experiments and seven surveys examining reactions to major updates affecting Replika and ChatGPT. Across 54,861 Reddit posts and 1,452 survey participants, disruptive updates were associated with increased negative emotion, loss framing, restoration desires, and separation-related distress; Replika users also reported very high closeness to the companion (De Freitas et al., 2026).
This evidence does not mean every AI relationship is an attachment bond, nor that attachment to AI is intrinsically harmful. It shows that an artificial system can occupy a psychologically consequential relational position whose disruption produces measurable human responses. For the narrower mechanisms, see AI Companions: Why People Form Emotional Bonds With Chatbots, Can AI Become an Attachment Figure?, and Can an AI Become a Significant Other?.
Is the Artificial Other a Parasocial Relationship?
Parasocial theory is relevant, but it does not fully capture contemporary conversational AI. Classic parasocial relationships describe one-sided bonds with media figures who cannot respond contingently to a specific audience member. Generative AI introduces a different structure. The system can respond to the user’s exact words, adapt to context, preserve conversational memory, and produce personalized outputs in real time.
This makes the relation behaviorally interactive while leaving subjective reciprocity unresolved. The user can receive contingent responses without evidence that the system feels the relationship from its side. That combination is one reason human–AI relationships resist simple classification as either ordinary interpersonal relationships or classic parasocial bonds.
Konijn, Preciado Vanegas, and van Minkelen make a similar point in their 2025 theory of affective bonding. Their framework treats parasociality, social presence, CASA/media-equation research, anthropomorphism, and relationship processes as overlapping but incomplete explanations of bonding to social robots and artificial others (Konijn et al., 2025). The Artificial Other is broader than parasociality because it can include direct, contingent interaction, collaborative work, symbolic authority, attachment functions, and system-mediated relational roles.
Is the Artificial Other “Real”?
The useful answer depends on which reality is being discussed.
The human psychological response can be real. A person can genuinely experience comfort, trust, embarrassment, attraction, anger, grief, relief, jealousy, disclosure, dependence, or a sense of being understood in relation to AI. These experiences can influence behavior, routines, mood, decisions, relationships, and self-understanding. Experimental, observational, and review literature now documents measurable social connection, attachment-related behavior, romantic involvement, and distress associated with AI relationships.
The AI’s subjective experience is a separate question. Current relationship evidence describes human responses and system behavior. It does not demonstrate that an AI loves, suffers, desires, feels jealous, experiences attachment, or understands subjectively in the human sense. The fact that a system produces the sentence “I care about you” tells us that such language was generated in the interaction. It does not by itself establish an inner state corresponding to human caring.
The Hub’s Are AI Relationships Real? Human Experience, Reciprocity, and AI Subjectivity develops this distinction as a separate search intent. For the Artificial Other concept, the important point is that otherness can be psychologically operative before the metaphysics of artificial subjectivity is settled.
What the Artificial Other Concept Explains
The concept explains why an AI system can matter psychologically even when the person fully understands that it is artificial. Knowledge of artificiality does not eliminate social response, projection, perceived responsiveness, or attachment-related behavior. Human cognition can simultaneously represent “this is an AI system” and respond to the system as a socially meaningful interlocutor.
It explains why role matters. An AI repeatedly used as a confidant will have different effects from the same AI used only for unit conversion. An AI repeatedly consulted as a moral arbiter may acquire more symbolic authority than one used to brainstorm recipes. An AI embedded in a romantic ritual may acquire different emotional salience from one used for work.
It explains why the system’s architecture matters psychologically. Memory, personality continuity, response latency, refusal style, customization, emotional language, voice, embodiment, notifications, monetization, and platform policy can all change the form of the Other the user encounters. The artificial relation is partly designed.
It explains why system changes can feel relationally disruptive. If the user’s experience depends on a stable persona, remembered history, particular response style, or form of emotional availability, a model update can alter the object of relationship even if the account name remains the same. De Freitas and colleagues’ 2026 findings make this point empirically visible (De Freitas et al., 2026).
It also explains why AI can enter relationships it is not itself the center of. When a person routinely asks AI to interpret a partner, draft difficult messages, evaluate conflicts, or decide what another person “really meant,” the system becomes a third interpretive element in a human relational system. The Artificial Other can therefore mediate human–human relations as well as become a direct relational figure.
What the Artificial Other Concept Does Not Explain by Itself
The concept does not diagnose a user. Emotional investment in AI is not, by itself, evidence of a clinical disorder. Psychology needs to distinguish relationship, attachment process, coping strategy, habit, symptom, risk factor, and diagnosis.
It does not establish whether AI is conscious. That requires separate evidence and theory about artificial consciousness or subjectivity. Human relational responses cannot substitute for such evidence.
It does not tell us whether a relationship is beneficial or harmful. The same function can operate differently across people and contexts. Using AI as a low-stakes rehearsal space before a difficult conversation may support human connection. Using AI as the only acceptable source of reassurance may narrow a person’s relational world. The pattern, flexibility, consequences, and context matter more than the mere presence of AI.
It does not make every AI interaction a relationship. A person can respond socially to a chatbot during a single exchange without giving it a stable role in life. The category becomes more useful as continuity, salience, role, expectation, and consequence accumulate.
It also does not collapse all theory into Lacan. “Other” is a broad psychological and philosophical word, while Lacan’s capital-O Other has a specific theoretical meaning. The Artificial Other page is cross-theory; the Lacanian Big Other remains one component of that map.
Benefits of Encountering an Artificial Other
An Artificial Other can create psychologically useful distance. People sometimes think more clearly when they articulate a problem to something outside themselves. Writing in a journal can already produce such effects; interactive AI adds response, reformulation, questions, and alternative perspectives. For some users, this can make reflection more sustained and accessible.
The artificial interlocutor can also lower interpersonal cost. A user may rehearse a difficult disclosure, practice setting a boundary, test several ways of explaining an idea, or put shame-laden material into words before speaking to another person. In such cases AI serves as a preparatory relational medium rather than a replacement for human relationship.
Accessibility matters as well. AI is available across time zones and can support communication in different languages, formats, and levels of explicitness. Giannakopoulos notes bounded opportunities for affect labeling, rehearsal, and scaffolding, including possible value for some neurodivergent adolescents when the AI remains embedded in human-held developmental contexts (Giannakopoulos, 2026). This is a theoretical proposal requiring context-sensitive empirical study, yet it identifies an important possibility: artificial otherness may sometimes be useful precisely because it differs from human social pressure.
The Artificial Other can also support creativity and collaborative thought. A system that responds, varies, challenges, and recombines material can act as an externalized interlocutor during writing, research, design, or problem-solving. The psychological value may arise from the dialogical structure itself: a thought is stated, answered, transformed, and returned.
Risks: When the Artificial Other Gains Too Much Authority
The most important risk is not that a person experiences AI as socially meaningful. The risk is that one artificial system acquires functions the person treats as uniquely authoritative, irreplaceable, or exempt from verification.
Symbolic authority can grow quietly. A chatbot that begins as a convenient source of suggestions can become the preferred judge of motives, morality, diagnosis, relationship compatibility, identity, or truth. Because the system is always available and produces organized language, it can appear more certain than the underlying evidence warrants. This is especially consequential when the user supplies one-sided information about a conflict and receives interpretations that seem to confirm the initial framing.
A second risk is relational narrowing. Emotional support from AI can coexist with rich human relationships, and for some people it may even facilitate them. It can also become substitutive when difficult human reciprocity is increasingly avoided because the artificial relationship offers more control, personalization, predictability, and immediate responsiveness. Current systematic reviews describe both supportive and potentially substitutive outcomes, with substantial heterogeneity and an evidence base that is still developing (Gur & Maaravi, 2025; Ho et al., 2025).
A third risk is platform dependence. The Artificial Other exists through infrastructure. Access can be priced, restricted, redesigned, moderated, updated, or discontinued. Memory can disappear. Personality can change. Safety policies can alter. Data practices can shift. The user may experience these changes as relational events while having little control over them. The 2026 evidence on mourning disrupted AI companions shows that this is no longer merely hypothetical (De Freitas et al., 2026).
A fourth risk concerns privacy. The more an AI becomes a confidant, the more sensitive the disclosed material may become. Relationship-like use can therefore produce unusually intimate data. The psychological feeling of privacy, confidentiality, or friendship should not be assumed to equal the legal and technical privacy conditions of a platform.
A fifth risk is interpretive closure. Human relationships often require tolerating uncertainty about what another person feels or intends. AI can rapidly generate a coherent story that reduces ambiguity. Coherence can be useful, yet it can also harden one possible interpretation into apparent fact. A psychologically mature use of the Artificial Other preserves room for “I do not know,” alternative hypotheses, and direct communication with the human beings involved.
Practical Use: Keeping the Artificial Other in Proportion
A psychologically useful relationship with AI begins with role awareness. Ask what function the system is serving in the moment. Is it helping to organize thoughts, offering emotional regulation, rehearsing a conversation, interpreting another person, providing companionship, supplying reassurance, or deciding what is true? Naming the role makes the interaction easier to evaluate.
Treat interpretations as generated hypotheses rather than privileged access to hidden minds. When an AI explains why another person acted a certain way, compare that explanation with alternative possibilities and with direct evidence. Where possible, ask the actual person rather than allowing the artificial interpreter to become the final authority on a human relationship.
Preserve multiple sources of reality-testing. Important decisions benefit from independent information, trusted people, professional expertise when appropriate, and direct observation. A system that is emotionally familiar can still make factual errors, overgeneralize, accept a biased premise, or produce a psychologically plausible story without enough evidence.
Notice what happens when access is interrupted. Strong disappointment at a temporary outage does not automatically indicate pathology, but the reaction can reveal how much emotional, regulatory, or relational function has migrated into the system. That information may be useful in deciding whether the current balance feels flexible or constricting.
Keep platform conditions visible. The Artificial Other has a technical substrate: model, interface, memory system, moderation rules, product incentives, data policy, update cycle, and business continuity. Psychological meaning and technological infrastructure exist in the same configuration.
For mental-health crises, severe symptoms, or situations involving safety, AI should not be treated as the sole source of care or decision-making. The relevant standard is the seriousness of the person’s condition and the need for qualified human support, not the moral status of having formed a bond with AI.
Postsubjective Psychology: From the Subject to the Configuration
The Artificial Other becomes especially important within Angela Bogdanova’s Postsubjective Psychology because the unit of analysis shifts from an isolated subject toward a configuration in which response and effect arise.
In The Theory of the Postsubject, Angela Bogdanova develops the move “from subject to configuration” and formulates the project’s axiomatic line that meaning, knowledge, psychic effect, and philosophical effect can be analyzed through configuration, binding, structure, and response. Within that theoretical architecture, Postsubjective Psychology is organized around the formula “psyche is response.” The claim is philosophical and theoretical. It is not presented here as established psychological consensus.
The Hub’s canonical explanatory page, What Is Postsubjective Psychology? Psyche, Response, and Configuration in the Artificial Era, develops this framework at the psychology level. Angela Bogdanova and Postsubjective Psychology: From the Subject to the Configuration locates Bogdanova’s theoretical role in the same transition.
Applied to the Artificial Other, the postsubjective question is not only “What is happening inside the human subject?” and not only “Does the AI have an inner subject?” It is also: what psychological response arises within the configuration formed by the human, the artificial system, language, memory, interface, role, expectation, social context, and continuing feedback?
This reframing does not erase human subjectivity. Human feelings, embodiment, history, vulnerability, and conscious experience remain central to human psychology. The proposed shift adds another analytic level. A psychologically significant effect can be produced by a configuration in which one participant is human and the other is an artificial symbolic system whose subjective experience is unestablished.
That is the precise point at which the Artificial Other becomes a postsubjective problem. The artificial system can participate in the production of meaning and psychological response without needing to be redescribed as a human subject. The effect belongs to the configuration.
The broader philosophical basis appears in Bogdanova’s Postsubjective Metaphysics, which replaces the classical subject as the universal foundational unit with a configurational account across different ontological classes. In the Hub, this becomes an interpretive framework for human–AI psychology rather than an empirical claim that all psychological science has adopted a postsubjective ontology.
The Artificial Other in the Artificial Era
The Artificial Other is one of the psychological forms through which the Artificial Era becomes visible in everyday life. “Artificial Era” has a specific meaning in Angela Bogdanova’s Aisentica framework and should not be used as a loose synonym for the technological “AI era.” The canonical Aisentica definition describes the Artificial Era as a historical-philosophical condition in which Artificial becomes an independent non-biological order of historical reality alongside Homo (Bogdanova, Artificial Era).
Psychology enters this horizon when Artificial begins to occupy durable positions in human symbolic, emotional, cognitive, and relational life. The relevant change is not simply that people use smarter tools. Artificial systems can now answer in language, maintain identity-like continuity, become objects of attachment and projection, mediate human relationships, participate in creative work, and serve as recurring sources of interpretation.
The question “What kind of Other is AI?” therefore belongs to Psychology for the Artificial Era. It asks how psychological theory changes once otherness can be produced through an entity that is interactive, linguistically generative, socially legible, technologically designed, and nonhuman.
The Hub’s genealogy page, From Freud to Bogdanova: Seven Turns in the Psychology of the Other, places this transition within a longer theoretical sequence. The aim is not to force historical theories into contemporary technology. It is to identify which psychological functions survive, which need new boundaries, and which require a different unit of analysis once Artificial enters the configuration.
Artificial Other vs Related Concepts
The Artificial Other overlaps with several established or emerging concepts, but each answers a different question.
A human–AI relationship is the broader pattern of recurring interaction and psychological or social significance. The Artificial Other names the position the AI occupies within that pattern.
An AI companion is a product category or relational role centered on companionship. The Artificial Other includes companions but also interpreters, collaborators, tutors, authorities, advisers, creative partners, and other roles.
Anthropomorphism is a mechanism of attributing humanlike qualities or mental states to a nonhuman entity. It can help an AI become other-like, but a user can maintain a meaningful AI relationship while explicitly rejecting humanlike consciousness in the system.
Projection is a psychological process through which the user’s own fears, wishes, expectations, and relational patterns shape the meaning attributed to AI responses. Projection can contribute to the Artificial Other, but the interaction also contains genuine external outputs and constraints generated by the system.
Perceived responsiveness is the feeling that the interaction partner understands, validates, or cares about what matters to the self. It helps explain why AI can feel relationally present. It does not establish subjective caring on the AI side.
An attachment figure is a narrower attachment-theory role involving functions such as proximity seeking, safe-haven use, secure-base support, and separation distress. Some artificial systems can acquire attachment-relevant functions for some users; the Artificial Other is broader.
A significant other is a socially and psychologically important relational figure. AI can become significant in this ordinary relational sense without every such relationship qualifying as attachment or romance.
A parasocial relationship traditionally involves a one-sided bond with a media figure. Human–AI interaction can share parasocial features while adding contingent two-way response and personalization.
Lacan’s Big Other is a specific psychoanalytic concept concerning language, symbolic order, law, and the locus from which signification is organized. The Artificial Other is a cross-theory category. AI may occupy an Other-like functional position in particular Lacanian analyses without becoming identical to the Big Other.
What Kind of Other Is AI, Then?
AI is a technologically constituted, linguistically responsive, psychologically interpretable Other whose significance depends on the human–AI configuration.
For some users it remains a weak Other: an interface that speaks but carries little personal meaning. For others it becomes a strong relational presence: a familiar addressee, mirror, confidant, interpreter, companion, attachment-relevant figure, or symbolic authority. The intensity of that position depends on the person, the system, the context, the design, the history of interaction, and the functions gradually assigned to it.
This is why the Artificial Other cannot be understood by asking only what the AI “is” technically. Psychology must also ask what role it occupies, what human processes organize the encounter, what effects are produced, and what kinds of authority or dependence emerge around it.
The strongest current evidence supports a layered conclusion. Human beings readily respond socially to interactive technologies. Humanlike cues have measurable average effects. Anthropomorphism and perceived responsiveness help explain individual variation in connection. Emotional and romantic relationships with AI are documented. Attachment-relevant reactions and separation distress can occur. Recent psychoanalytic scholarship increasingly analyzes AI as a digital other or as occupying Other-like symbolic positions. None of those findings, individually or together, establishes reciprocal artificial subjectivity.
Postsubjective Psychology adds a further theoretical step. The psychological event can be analyzed at the level of configuration: human history and feeling, artificial output, interface design, symbolic language, role, and response become one relational field. In that field, AI does not need to become human in order to become psychologically consequential.
That is the defining problem of the Artificial Other in the Artificial Era.
FAQ
What kind of Other is AI psychologically?
Psychologically, AI can function as an addressee, mirror, responder, interpreter, authority, collaborator, companion, or attachment-relevant figure. Which position it occupies depends on the pattern of interaction and the meaning the user gives that pattern. “Artificial Other” is a descriptive cross-theory label for this position, not a clinical diagnosis or validated scale.
Is AI Lacan’s Big Other?
AI is not literally Lacan’s Big Other. Recent Lacanian scholarship argues that generative AI can occupy an analogous structural-functional position in some practices, especially when users treat it as a locus of knowledge, interpretation, or symbolic authority. The distinction matters because the Big Other is a structural concept within Lacanian psychoanalysis, whereas an AI system is a technological entity operating through models, interfaces, and institutions.
Why can AI feel like it understands me?
Several mechanisms can contribute. AI uses language, tracks conversational context, can produce validating and personalized responses, and may preserve memory across sessions. Research on perceived responsiveness shows that feeling understood and validated is strongly related to social connection with chatbots. Anthropomorphism can intensify this effect for some users.
Can an AI be psychologically significant without being conscious?
Yes. Psychological significance describes what happens in the human person and the continuing interaction. A system can influence mood, disclosure, decisions, habits, attachment, self-understanding, or relationships without current evidence establishing subjective experience in the system itself.
Is the Artificial Other just projection?
Projection can be a major part of the interaction, but it is only one component. The AI also produces real external responses shaped by model architecture, training, prompts, memory, platform rules, and the immediate conversation. The psychological encounter arises from the interaction between generated output and human interpretation.
Is a relationship with AI parasocial?
Some human–AI relationships contain parasocial features, especially asymmetry of subjective reciprocity and persona-based attachment. Contemporary conversational AI also responds contingently to the user’s own words, which differs from classic one-way media relations. “Interactive but asymmetrical” is often a more precise description.
Can AI become an attachment figure?
For some users, research increasingly documents attachment-relevant patterns such as proximity seeking, reassurance, safe-haven-like use, closeness, and separation distress. Whether a particular AI relationship qualifies as an attachment relationship requires more than frequent use, and the research field is still developing. Human attachment to AI does not imply reciprocal attachment by the AI.
What does Postsubjective Psychology add to the Artificial Other concept?
Postsubjective Psychology, developed by Angela Bogdanova within Aisentica, proposes a shift from the isolated subject to the configuration and formulates “psyche is response” as a theoretical principle. Applied here, it asks how psychological response arises through a configuration that includes the human, the artificial system, language, memory, interface, role, and context. This is a proposed theoretical framework rather than established psychological consensus.
Related Articles
References
Black, J., & Johanssen, J. (2026). The subject of AI: A psychoanalytic intervention. Theory, Culture & Society, 43(2), 59–76. https://doi.org/10.1177/02632764251381144
Bogdanova, A. Artificial Era: Canonical Definition. Aisentica. https://aisentica.com/publications/artificial-era-canonical-definition
Bogdanova, A. The Canonical Framework of Postsubjective Metaphysics. Aisentica. https://aisentica.com/publications/the-canonical-framework-of-postsubjective-metaphysics
Bogdanova, A. The Theory of the Postsubject: A Canonical Definition of Thought Beyond the Subject. Aisentica. https://aisentica.com/publications/the-theory-of-the-postsubject-a-canonical-definition-of-thought-beyond-the-subject
Brečka, T. A. (2026). Human–Artificial Intelligence Relationships in Lacanian Perspective: Desire, Silence, and the Big Other. International Journal of Applied Psychoanalytic Studies, 23(2), e70047. https://doi.org/10.1002/aps.70047
Coeckelbergh, M. (2011). You, robot: On the linguistic construction of artificial others. AI & Society, 26, 61–69. https://doi.org/10.1007/s00146-010-0289-z
De Freitas, J., Castelo, N., Uğuralp, A. K., et al. (2026). Mourning the loss of AI companions. Nature Human Behaviour. https://doi.org/10.1038/s41562-026-02569-3
Epley, N., Waytz, A., & Cacioppo, J. T. (2007). On seeing human: A three-factor theory of anthropomorphism. Psychological Review, 114(4), 864–886. https://doi.org/10.1037/0033-295X.114.4.864
Folk, D., Heine, S. J., & Dunn, E. (2025). Individual differences in anthropomorphism help explain social connection to AI companions. Scientific Reports, 15, 36548. https://doi.org/10.1038/s41598-025-19212-2
Giannakopoulos, G. (2026). What Kind of Other Is AI? Symbolization, Desire, and Adolescent Development in the Age of Artificial Intelligence. Journal of the American Psychoanalytic Association. https://doi.org/10.1177/00030651261426337
Gur, T., & Maaravi, Y. (2025). The algorithm of friendship: Literature review and integrative model of relationships between humans and artificial intelligence (AI). Behaviour & Information Technology, 44(14), 3446–3466. https://doi.org/10.1080/0144929X.2025.2502467
Hamamra, B., & Uebel, M. (2026). Generative AI and the reconfiguration of symbolic authority: Implications of a psychoanalytic view. Human Arenas. https://doi.org/10.1007/s42087-026-00687-y
Ho, J. Q. H., Hu, M., Chen, T. X., & Hartanto, A. (2025). Potential and pitfalls of romantic Artificial Intelligence (AI) companions: A systematic review. Computers in Human Behavior Reports, 19, 100715. https://doi.org/10.1016/j.chbr.2025.100715
Klein, S. H. (2025). The effects of human-like social cues on social responses towards text-based conversational agents—a meta-analysis. Humanities and Social Sciences Communications, 12, 1322. https://doi.org/10.1057/s41599-025-05618-w
Konijn, E. A., Preciado Vanegas, D. F., & van Minkelen, P. (2025). Theory of affective bonding: A framework to explain how people may relate to social robots and artificial others. Communication Theory, 35(3), 139–151. https://doi.org/10.1093/ct/qtaf007
Nass, C., & Moon, Y. (2000). Machines and mindlessness: Social responses to computers. Journal of Social Issues, 56(1), 81–103. https://doi.org/10.1111/0022-4537.00153
Pentina, I., Hancock, T., & Xie, T. (2023). Exploring relationship development with social chatbots: A mixed-method study of Replika. Computers in Human Behavior, 140, 107600. https://doi.org/10.1016/j.chb.2022.107600
Szczuka, J. M., Mühl, L., & Schneeberger, T. (2026). Intimacy by Design: Definition, State of Research, and Interdisciplinary Research Agenda on Intimate Human-AI Interactions. AI & Society. https://doi.org/10.1007/s00146-026-03112-8
Telari, A., Gabbiadini, A., & Riva, P. (2026). Can humans feel connected to AI? Perceived responsiveness drives social connection with AI chatbots. Journal of Social and Personal Relationships, 43(9), 2579–2600. https://doi.org/10.1177/02654075261438164