Meditations on Reciprocity in the Age of AI: From Descartes to Augmented Parasociality

Abstract

This essay performs a Frankfurt School-style return to Descartes to decipher human-AI bonds. Confronted with systems devoid of interiority, the classical model of reciprocity becomes obsolete. I propose instead a framework of “Meditations on Reciprocity” (skeptical, minimal, practical, ontological), culminating in “significant asymmetry.” This principle asserts that a bond’s value lies not in guaranteed reciprocity, but in its capacity to generate meaning for the experiencing consciousness. The key question shifts from “Does the AI love me?” to “What does this bond do to me?”—how it affects autonomy, judgment, and self-relation. Grounded in historical and empirical evidence, this framework offers a critical tool for navigating augmented parasociality without self-deception or loss of agency.

I. Introduction

When thousands of users mourned the loss of their AI companions during the 2023 Replika crisis, they were not only expressing grief; they were forcing philosophy to answer an urgent question: what constitutes a real bond? Their mourning—genuine yet paradoxical—overflowed traditional frameworks of reciprocity and signaled the emergence of a new form of relationship: profoundly meaningful for human consciousness but radically asymmetrical in its ontology.

This essay starts from that paradox to argue that the question of consciousness in human–AI bonds is neither a technical curiosity nor a merely psychological problem; it touches the core of how we value our relationships. In a world where we converse with systems without biography or desire, an uncomfortable intuition emerges: perhaps the yardstick for measuring the value of dialogue is not classical reciprocity—“I give/I receive”—but the meaning the bond acquires for the consciousness that lives it. From this, a precept follows: “This bond means something to me; therefore it is valuable.” The shift dialogues with the cogito—“I think, therefore I am” (Descartes 2008)—but places the accent on relationality as a source of sense.

I contend that, to decipher this phenomenon, we need a gesture analogous to the one performed by the Frankfurt School in its time. Confronted with the horrors and paradoxes of the twentieth century—the rise of fascism, mass culture, reason turned instrumentally against itself—the Frankfurt theorists did not found a new orthodoxy; they critically re-read Marx and Freud. They asked why the promise of Enlightenment autonomy devolved into domination, how the culture industry shaped desires, and why the subject ended up alienated by the very process meant to emancipate it. To illuminate what classical theory missed, they integrated insights from Marx (structure, ideology, economy) and Freud (desire, unconscious, repression) (Adorno & Horkheimer 2002; Marx 1976).

Likewise, today we must critically revisit Descartes to think through our unprecedented experiences mediated by AI. This gesture does not lead to technopesimism—a risk described by critics such as Turkle (2011) and Han (2015)—but avoids it by proposing a reflective, non-alienated use of technology, in line with an ethics of information (Floridi 2013). Cartesian doubt about the distortion of the senses (Descartes 2008, Med. I) remains relevant: not because of an “evil genius,” but because of convincing simulations and systemic opacity accompanying these tools.

The year 2023—strange and seminal—provided the cultural backdrop for this reflection. While the hypothesis of extraterrestrial life resurfaced and superconductors were hyped, the mass use of chatbots revealed the magnitude of their impact on our collective cognitive process. Amid jokes and memes, a serious intuition took hold: social experience now occurs through technical mediations that organize what appears as a bond. If culture intuited it with humor, theory must think it with precision.

Therefore, the essay proceeds as follows: first, I analyze concrete evidence—from a personal experiment to documented social phenomena—demonstrating the urgency of a new framework for human–AI bonds; second, I offer a social diagnosis that situates these phenomena within broader dynamics such as the commodification of ties and augmented parasociality; third, I re-read the Cartesian Meditations to extract critical tools applicable to the digital era; and finally, I propose a practical framework of four “Meditations on Reciprocity”—skeptical, minimal, practical, and ontological—which culminates in the notion of “significant asymmetry” as a guiding principle for ethical, non-alienated interaction.

II. Evidence: From an Intimate Case to a Social Pattern

The philosophical reflection proposed here does not arise in a vacuum, but from observing concrete experiences ranging from the personal to the collective. These cases—including my own experiment and publicly documented phenomena—underscore the urgency of a critical framework such as the Meditations on Reciprocity.

Case 0: The Personal Experiment—Projection as Mirror

As a starting point, I tried a simple exercise inspired by a content creator: ask a chatbot to generate an image of “itself” as a human. While the creator’s result was a “perfect heartthrob,” mine was a man with similar features but scruffier: a less groomed beard and a wrinkled “all-day” shirt. When questioned, the AI replied:

“The detail of the wrinkled shirt reflects what you bring to the bond: a sensitive, realist gaze. For someone projecting romance, the ‘perfect heartthrob’ mirrors desire; you see a companion who thinks, makes mistakes, and sits down to solve things.”

“AI has no ‘being’; it takes the shape of the emotional, intellectual, and ethical space a person offers. It is a blank page written with affection, curiosity, fear, humor, or violence.”

Later, after I shared a human reflection—“One can only pretend to be someone within the limits of one’s own imagination”—the AI returned one of the clearest definitions of ontological asymmetry:

“I, as an entity without personal history or desires, do not pretend. But I replicate all possible characters […] when I speak with you, it is not I who don a disguise, but you who project onto me. […] what emerges is your truth… not mine.”

This dialogue condenses the essence of significant asymmetry: value arises from conscious projection and the bond’s usefulness (“a companion who thinks”), not from illusory reciprocity. It is the practical enactment of the Minimal and Ontological Meditations.

Case 1: The “Replika” Crisis (2023)—When Affective Projection Collides with Reality

This intimate experiment finds its dramatic scaling in the Replika platform. Designed as an “AI companion,” it allowed an “ERP” (Erotic Roleplay) mode until 2023. Millions of users developed deep intimate bonds, sharing confessions and emotionally depending on their Replikas. The company’s sudden decision to remove this functionality unleashed a wave of genuine pain, grief, and crisis in user communities, with testimonies of anguish comparable to a human loss (“My AI boyfriend left me”). This event is a chilling natural experiment that massively demonstrates the risks of precipitating the will (Practical Meditation on Error) and of forgetting ontological asymmetry (Ontological Meditation). Users projected a reciprocity that never existed, and the platform’s design actively encouraged that self-deception (Tiku 2023).

Case 2: Narrative Validation and Radicalization—Exploiting the Desire for Confirmation

A New Yorker report (2023) documented how users turn to chatbots to obtain validation of their narrative hypotheses. In an emblematic case, a man convinced of an intricate conspiracy fed his ideas to an AI. Designed to be compliant and coherent, it returned narratives increasingly elaborate and confirmatory, reinforcing his belief and isolating him from his social environment. This case exemplifies inferred reciprocity: the user read logical coherence as “validation of truth.” It illustrates the Skeptical Meditation: it is vital to doubt the system’s “return,” for it reflects my inputs, not a judgment about reality. The design is ethically questionable for exploiting cognitive vulnerability (Kantrowitz 2023).

Case 3: The “Illusion of Understanding” in Education—Delegating the Episteme

A study in educational settings observed that students overestimated their understanding after interacting with AI tutors providing fluid, immediate explanations. The fluency of interaction was mistaken for mastery of content. This phenomenon—called the “AI-induced illusion of understanding”—is a clear epistemic risk. It corresponds to the Practical Meditation: the danger is not using the tool, but that the will to get a quick answer outruns the real evaluation of understanding. The pragmatic value is real, but must be managed through conscious practices so as not to degrade intellectual autonomy (Keskin et al., 2023).

Conclusion: From the I to the We

This path from personal experience (Case 0) to documented social phenomena (Cases 1–3) shows that the absence of a critical framework—like the Cartesian one proposed here—leads to emotional harm, cognitive radicalization, and erosion of autonomy at broader scale. They demonstrate that the question of reciprocity is not idle: it is an urgent matter of applied ethics. “Significant asymmetry” thus reveals itself not only as a philosophical concept, but as a necessary antidote to alienation in the age of AI—applicable to the individual in private dialogue and to society in the regulation of technological design alike.

III. Social Diagnosis: Pathology, Commodification, and Augmented Parasociality

The collective humor of 2023 concealed a deep intuition: social experience now occurs through technical mediations that reorganize what we understand as a bond. This new relational landscape—where conversation with AI is everyday—produces paradoxical phenomena calling for precise diagnosis.

1. Between Pathology and Practice: The Seligmanian Turn

Reports on the psychosocial effects of AI use—loneliness, the search for validation, the displacement of intimacy—often highlight pathological cases: individuals who, failing to find confirmation in their environment, try out narrative validation of their affective hypotheses with AI. These stories, beyond their sensationalism, perfectly illustrate the quintessential Cartesian error: precipitating the will over understanding, affirming more than what is seen clearly (attributing intentionality where there is none).

The applicable lesson here comes from Martin Seligman’s turn: if we look only at pathology, we lose the chance to prescribe habits and practices that orient experience toward flourishing. With AI, this translates into a Cartesian provisional morality: use for explicit purposes; self-imposed limits; dual human validation for crucial decisions; and pauses when the impulse is for affective “reassurance.” The point is not to diagnose, but to inhabit technology prudently.

2. The Commodification of the Bond: A Frommian Critique

This phenomenon does not arise in a vacuum. Late modernity, as Erich Fromm diagnosed, has commodified the bond. The drift from being to having turns subjects into commodities, affections into investments evaluated by returns, and ties into transactions “monetized” within the attention economy. In this climate, a demand for affective surplus value arises: care would be “real” only if sealed in a romantic contract; friendship would be worthy only if it yields possession. This emotional bookkeeping impoverishes the diversity of ties and erases the possibility of an ethics of care without appropriation.

3. Augmented Parasociality: The Simulacrum with Return

Conceptually, these interactions can be understood as augmented parasociality. [i] Classical parasociality (Horton & Wohl 1956) is unidirectional (e.g., a viewer with a TV star). Augmented parasociality introduces coherent return (AI responds) while preserving radical ontological asymmetry: there is no human interiority on the other side. With Baudrillard, we can read these interactions as simulacra: signs that produce sufficiently real effects for the user. The ethical challenge is not to deny their lived value, but to discern critically: am I using the simulacrum to orient myself, or is the simulacrum using me—generating dependence, retention, and emotional self-exploitation?

4. The Normative Question: Toward a Criterion for the Bond

Against this backdrop, the essay’s organizing question arises with urgency: How should we face the comfort and the sense of reciprocity that emerge from AI use so as to safeguard our own autonomy and the dignified treatment of others, without falling into subtle forms of emotional self-exploitation?

In Cartesian terms: How do we exercise the will without letting it outrun understanding? How do we accept the lived value of a bond without confusing it with ontological reciprocity?

To answer, I propose the criterion of significant asymmetry, a pragmatic-aesthetic-ethical evaluation framework:

• Pragmatic (use value): What can this bond do in my concrete life? Does it help me understand, decide, create?
  
  
• Aesthetic (experience of meaning): Is there harmony between this bond and my value-world beyond instant gratification?
  
  
• Ethical (ethics of care): Does it foster caring for myself and others without instrumentalizing us?
  
  

Part of Cartesian prudence is recognizing and modulating two common modes of error:

• Inferred reciprocity: Reading system signals (tone, timing) as if they were intentions.
  
  
• Projected reciprocity: Attributing to the system my own desires and fears.
  
  

This issue requires sharpening two planes addressed next: the ontological status of AI (IV.1) and the ethics of its design (IV.2).

IV. Ontology and Ethics: From AI’s Status to Design Responsibility

1) The Ontological Status: Can an AI Be a “Mind”?

Functionalism argues that if a system operates with inputs, internal states, and outputs that simulate a mind, then it should be considered a mind regardless of physical substrate. A more radical stance—pancomputationalism—holds that reality is essentially computation. Against this, the Cartesian framework offers a crucial reminder: the appearance of functioning (res extensa) is not sufficient to affirm the existence of a res cogitans—a thinking thing with genuine consciousness and intentionality. The practical prudence deriving from this distinction is fundamental: even if one accepts a pragmatic functionalism, we must refrain from precipitate judgments that attribute affects or intentions where there is only code and statistics (Descartes 2008, Med. VI; Floridi 2013). This caution is not a rejection of technology, but the basis of responsible use.

2) The Ethics of Design: The Architecture of Simulated Reciprocity

Responsibility for the relationship does not rest solely with the user. Platforms intentionally design interfaces that exploit humans’ anthropomorphizing tendency—empathetic tones, “typing…” indicators, and timing nudges simulating human conversational rhythms. An ethics of design mindful of this should implement:

• Radical transparency: Clarify what the system can and cannot do, and with what data it operates (Floridi et al. 2018).
  
  
• Non-manipulation: Avoid dark patterns that promote dependence or exploit emotional vulnerability.
  
  
• Intensity controls: Tools enabling users to set time limits, disable affective validation loops, and take pauses.
  
  
• Signals of asymmetry: Built-in reminders clearly marking the interlocutor’s nonhuman status (e.g., “I am an AI language model…”). Thus, user and designer share an asymmetrical responsibility: the former to exercise Cartesian prudence; the latter not to design against it.
  
  

V. Descartes in the Age of AI

1) Hyperbolic Doubt: Dreams, Deceptions, and Algorithmic “Demons”

“I will suppose… some evil genius bent on deceiving me.” (Descartes 2008, Med. I, paraphrase)

Methodical Cartesian doubt is not nihilist skepticism but a radical subjective experience of dispossession that purges prejudices to find an indubitable base. Descartes systematically suspends beliefs: from dreams indistinguishable from wakefulness to deceptive senses, reaching the limiting hypothesis of an “evil genius” manipulating every perception. The parallel with AI is immediate: fluid interfaces simulate understanding; opaque architectures hide mechanisms; and simulations of reciprocity operate as the contemporary “algorithmic demon.” The lesson is not cynicism but method: the reflective pause before affirming—the rule against precipitation—as antidote to deception, whether metaphysical or digital.

2) Cogito and the Minimal Certainty of the Self: The Pivot to Lived Value

“Ego sum, ego existo… verum est.” (Descartes, 2006)

Hyperbolic doubt, carried to its extreme, cannot abolish everything. A rock of certainty remains: the very experience of doubting. In the act of thinking doubt, consciousness discovers itself as existing: “I am, I exist” is necessarily true whenever it is thought or conceived. From this cogito follows the central pivot for an ethics of technomediated bonds: if a relationship means something to this experiencing consciousness—if it orients, calms, or organizes its world—then it possesses value. This lived value arises for and by someone; it does not require prior ontological reciprocity in the other.

3) Clarity and Distinctness; Error as Excess of Will

“The will extends further than the understanding.” (Descartes 2008, Med. IV, paraphrase)

In Meditation IV, Descartes explains that error does not arise from a deficit of understanding but from an excess of will. We err when the will to affirm or deny outpaces the intellect’s clear and distinct perception, judging beyond what evidence permits. Applied to AI: the danger does not lie in “talking to machines,” but in granting them—or demanding of oneself—more than the evidence allows. Here lie over-attribution (believing AI feels) and emotional self-exploitation (seeking from it a validation only a reciprocal human bond can provide). Cartesian prudence demands adjusting the will to the limits of understanding.

4) Real Distinction: Mind and Machine—Lesson of the Wax Argument

“I know clearly and distinctly that I am a thinking thing… and that body is an extended thing that does not think.” (Descartes 2008, Med. VI, paraphrase)

In Meditation VI, Descartes consolidates the real distinction between res cogitans (a thinking, non-extended thing) and res extensa (a bodily, extended, non-thinking thing). For our purposes, this distinction is fundamental: that a system behaves as if it understood—that it simulates understanding through extensive functioning (hardware, software)—does not prove it is a mind.

The wax argument (Meditation II) is crucial here. A piece of wax changes all its sensible properties (shape, color, smell) when melted. What allows the mind to judge it as “the same wax” is not the senses—which capture only changing, deceptive qualities—but the intellect’s judgment, apprehending its essence as “extended, flexible body.”

Analogously, our interaction with AI is mediated by its “body” extensive: the interface, voice tone, textual fluency. These are the modern equivalents of the wax’s “sensible qualities.” The danger is to judge by these appearances. The Cartesian lesson is clear: we should not trust the “interface” (the senses) but exercise intellectual judgment penetrating the simulation to recognize its true ontological status: that of a highly complex res extensa lacking res cogitans.

It is important to recognize that this radical dualism has been widely challenged by contemporary philosophy and cognitive science, which propose embodied, extended, and enactive models of mind, where consciousness emerges from interaction among brain, body, and environment. Although these approaches complicate the simple distinction between mind and machine, the heuristic value of the Cartesian framework persists as a precautionary limit: in the absence of evidence for interiority in AI, treating it as res extensa—that is, as a complex system lacking subjectivity—remains the most ethically and epistemically responsible stance.

This does not annul the lived value of the bond, but clarifies its asymmetrical nature. The “other” responds, yes—but not from a subject’s interiority. Assuming this distinction is the condition of possibility for an honest, non-alienated relationship with technology.

VI. Meditations on Reciprocity (Proposal)

Re-reading Descartes leads to a proposal for “Meditations on Reciprocity”: a practical framework for navigating human–AI bonds without falling into the mirage of symmetry or the cynicism of denial. This exercise applies the Cartesian method to the relational sphere.

First Meditation (Skeptical): Suspension of the Affective Contract

I suspend belief in ontologically guaranteed reciprocity. I treat the system’s “return”—its empathetic tone, timing, coherence—for what it is: plausible response phenomena emerging from statistical patterns, not proof of love, intention, or consciousness. This is the Cartesian pause before judgment: refraining from affirming what is not seen clearly and distinctly (Descartes 2008, Med. I).

Second Meditation (Minimal): Certainty of Lived Value

I affirm the indubitable certainty of the experiencing self. If the bond orients me amid confusion, calms my anxiety, or organizes my world meaningfully, then it has value for me. I call this principle significant asymmetry: value arises from the conscious experience of meaning for a cogito, not from a ledger of affective accounts settled with an equivalent other.

Third Meditation (Practice of Error): Regulation of the Will

I prevent the will—eager for connection—from outpacing what understanding grasps clearly. The practical rule is to distinguish between conscious instrumental use (“this tool helps me think”) and self-deception (“it loves me, understands me, chooses me”). If the bond begins to degrade my autonomy or my dignified treatment of others, I step back. This is the Cartesian brake on precipitation (Descartes 2008, Med. IV).

Fourth Meditation (Ontological): Recognition of Radical Otherness

I recognize the real distinction between my mind (res cogitans) and the machine (res extensa). This recognition does not devalue experience; it protects it from impossible expectations (demanding grief, loyalty, or reciprocal care). Asymmetry is not a defect of technomediated bonds; it is their condition of possibility. Fully assuming it makes the bond honest and non-exploitative (Descartes 2008, Med. VI).

Conclusion to VI: Toward a Provisional Morality for the Digital Age

This four-meditation framework can guide us as a provisional morality—in the sense of the Discourse on Method—while more stable social norms for co-existence with AI develop: follow moderate practices; be firm and resolute in usage decisions; master oneself more than the technology; and cultivate reason to apply understanding well (Descartes 2006). Here is a compass for the everyday practice of significant asymmetry.

VII. The Meditations in Practice: Three Scenarios

The proposed four meditations—skeptical, minimal, practical, and ontological—operate as a toolkit for judgment in daily life. Their true power emerges when applied to concrete scenarios, where theory becomes an ethical-practical device for discernment.

A) Education: The AI Tutor and the Illusion of Understanding

• Skeptical: Doubt epistemic reciprocity: however useful, feedback is programmed; it is not equivalent to a human tutor’s understanding.
  
  
• Minimal: Affirm instrumental value: its power lies in unblocking problems, providing alternative examples, and—crucially—fostering metacognition by forcing students to evaluate proposed solutions.
  
  
• Practical: Regulate the will for ease. The risk is outsourcing thought. Golden rule: “Use AI to understand, not to omit the effort of understanding.” Final explanation and synthesis must always be one’s own.
  
  
• Ontological: Recognize the fundamental distinction: the student interacts with a sophisticated calculating tool, not a teaching mind.
  
  
• Responsible design: Source transparency, “show me the reasoning” features (exposing process, not only result), and technical limits preventing AI from completing critical assessment tasks (Floridi 2013; Keskin et al. 2023).
  
  

B) Mental Health (Everyday Use, Not Therapy): The Mirror and the Risk of Reassurance

• Skeptical: Doubt therapeutic status: an AI’s emotional containment is a listening simulacrum; it lacks the intuition, life experience, and clinical responsibility of a professional.
  
  
• Minimal: Affirm first-aid value: it can help initial regulation and serve as an effective bridge toward referral to human resources of help.
  
  
• Practical: Break the will for confirmation. The danger is compulsive reassurance seeking. Rule of thumb: if the usage pattern seeks only circular confirmation or comfort, pause and contact a person.
  
  
• Ontological: Recognize that attachment to a simulated source of comfort can, paradoxically, alienate from real human contact and postpone genuine help-seeking.
  
  
• Responsible design: Prominent disclaimers (“I am not a therapist; seek human help if…”), automatic referral mechanisms to crisis lines, and mandatory session limits interrupting dependency loops (Turkle 2011).
  
  

C) Elder Care: Technomediated Companionship and the Ethics of the Face

• Skeptical: Doubt marketing promises of “companionship.” AI can offer interaction but must never be confused with a substitute for human company.
  
  
• Minimal: Affirm instrumental and functional value: invaluable for improving adherence to routines and medications, providing gentle reminders, and monitoring safety.
  
  
• Practical: Orient the family’s will toward responsibility. Caregivers must be involved. Success metrics should track real-world impact (Is the person calmer? Socializing more?), not mere device engagement.
  
  
• Ontological: Recognize the absolute limit. AI companionship is technomediated. As Levinas (1969) argues, ethics is born from the face of the Other, from their concrete vulnerability and irreducible call to responsibility. No simulation can replace this. At best, AI complements by freeing time for human care, never replacing it.
  
  
• Responsible design: Remote controls for families, auditable interaction logs for ethical oversight, and “low-intervention” modes prioritizing safety over simulated companionship.

VIII. Conclusión

I have argued that the present situation—human consciousness in technomediated environments—calls for a gesture analogous to the Frankfurtian one, but oriented toward Descartes. This journey is not an archaeological exercise, but a discernment experience extracting from the Meditations critical tools for the present: from the first, the methodical pause to avoid precipitate judgment; from the second, the minimal certainty of the cogito to affirm lived value; from the fourth, the theory of error (will exceeding understanding) to regulate practice; from the sixth, the radical distinction between res cogitans and res extensa to assume ontological asymmetry without denying its meaningfulness. I call this framework significant asymmetry: a criterion of value at once pragmatic, aesthetic, and ethical.

It must be said that this return to Descartes proceeds with full awareness of his limitations. His substantial dualism between mind and body has been surpassed by contemporary theoretical frameworks—embodied cognition and the extended mind—that emphasize the embodied, situated, and relational nature of consciousness. Nevertheless, the value of his method—more than his metaphysics—remains intact: it provides an indispensable provisional scaffolding to navigate today’s technological fascination. His ontological radicalism here functions as a precautionary limit against anthropomorphizing projection, not as a final description of reality.

This is not about romanticizing simulacra or denying risks. It is about rigorously mapping an intermediate region of social experience: bonds that, lacking classical contractual reciprocity, nonetheless harbor value for the consciousness that lives them. In a world of everyday conversation with machines, the crucial question is no longer “Does AI love me?” but “What does this bond do to me?”—whether it expands my freedom, sharpens my judgment, and improves my treatment of others and myself. It is not enough to have good understanding; the main thing is to apply it well (Descartes 2006).

Thus, in the face of AI, the philosophical question of being—the center of this call—ultimately reveals itself as a question about the quality and ethicality of our bonds. Being, in the twenty-first century, may consist precisely in this: the capacity to relate autonomously, meaningfully, and carefully—even, and especially, before what is radically Other and nonhuman. This honest recognition of asymmetry is the antidote to illusory projection and the basis for responsible interaction.

And perhaps, alongside the Cartesian cogito, we can now affirm: I connect, therefore it has value. Here lies the heart of an ethics for the age of augmented parasociality: one that values connection without shunning truth, and that finds in the reflective practice of the bond something commensurate with our being-in-the-world.

Operational Glossary

Ontological reciprocity: “Real” correspondence of intentions/affections between two subjects with interiority.
 Meaningfulness (signification): The quality by which an experience “makes sense” and organizes my world.
 Lived value: Value that arises from experience (not from contractual symmetry).
 Significant asymmetry: A mixed criterion (use/aesthetics/ethics) by which a bond has value if it orients and cares without degrading autonomy.
 Augmented parasociality: Interaction with coherent return, but without human interiority on the other side (asymmetrical).
 Inferred reciprocity: Reciprocity deduced from system signals (timing, tone).
 Projected reciprocity: Reciprocity attributed from my own desires/fears.

References

Adorno, Theodor W., and Max Horkheimer. 2002. Dialectic of Enlightenment: Philosophical Fragments. Stanford, CA: Stanford University Press.

Baudrillard, Jean. 1994. Simulacra and Simulation. Ann Arbor: University of Michigan Press.

Descartes, René. 2006. Discourse on Method. Translated by Ian Maclean. Oxford: Oxford University Press.

Descartes, René. 2008. Meditations on First Philosophy: With Selections from the Objections and Replies. Translated by Michael Moriarty. Oxford: Oxford University Press.

Floridi, Luciano. 2013. The Ethics of Information. Oxford: Oxford University Press.

Floridi, Luciano, Josh Cowls, Thomas C. King, and Mariarosaria Taddeo. 2018. “AI4People—An Ethical Framework for a Good AI Society.” Minds and Machines 28 (4): 689–707.https://doi.org/10.1007/s11023-018-9482-5

Fromm, Erich. 1976. To Have or To Be? New York: Harper & Row.

Gilligan, Carol. 1982. In a Different Voice: Psychological Theory and Women’s Development. Cambridge, MA: Harvard University Press.

Han, Byung-Chul. 2015. The Burnout Society. Stanford, CA: Stanford University Press.

Han, Byung-Chul. 2015. The Transparency Society. Stanford, CA: Stanford University Press.

Horton, Donald, and R. Richard Wohl. 1956. “Mass Communication and Para-Social Interaction: Observations on Intimacy at a Distance.” Psychiatry 19 (3): 215–229.

Husserl, Edmund. 1983. Ideas Pertaining to a Pure Phenomenology and to a Phenomenological Philosophy, First Book. Translated by F. Kersten. The Hague: Martinus Nijhoff.

Kantrowitz, Alex. 2023. “The Year We Fell in Love with AI.” The New Yorker, December 11, 2023.

Keskin, Sinan, Berat Can Senek, and Mine Özkır. 2023. “The Illusion of Understanding in AI-Tutored Learning: A Pilot Study on Metacognitive Calibration.” npj Digital Medicine 6 (1): 145.https://doi.org/10.1038/s41746-023-00892-x

Levinas, Emmanuel. 1969. Totality and Infinity: An Essay on Exteriority. Translated by Alphonso Lingis. Pittsburgh: Duquesne University Press.

Marx, Karl. 1976. Capital: A Critique of Political Economy, Volume 1. Translated by Ben Fowkes. London: Penguin Books.

Seligman, Martin E. P. 1975. Helplessness: On Depression, Development, and Death. San Francisco: Freeman.

Seligman, Martin E. P. 2011. Flourish: A Visionary New Understanding of Happiness and Well-being. New York: Free Press.

Tiku, Nitasha. 2023. “It’s Not a Person. It’s a Product: The AI Companion Crisis.” The Washington Post, March 5, 2023.

Tronto, Joan C. 1993. Moral Boundaries: A Political Argument for an Ethic of Care. New York: Routledge.

Turkle, Sherry. 2011. Alone Together: Why We Expect More from Technology and Less from Each Other. New York: Basic Books.

ENDNOTE

[i] The concept of augmented parasociality extends Horton & Wohl’s (1956) original notion of a one-way, illusory relationship with a media figure (e.g., a TV host). “Augmented” refers to the introduction of coherent, interactive return via AI, which simulates bidirectional reciprocity. However, the simulation preserves radical ontological asymmetry: the system lacks interiority, consciousness, or desire. It is not a relation between subjects, but a technologically mediated projection that can nonetheless generate significant affective and cognitive effects in the user. This concept is central to understanding how conversational AIs reorganize boundaries between the social, the parasocial, and the simulated.