From Mnemosyne to the Synthetic Dead Mother? Exploring the Emotional Effects of Chatbot Use

by Joana Pizarro Bravo & Filipe Leão Miranda

In Greek mythology, desire for mastery over memory and thought was represented by the Muses, born from Zeus and Mnemosyne, the goddess of memory. To the ancient Greeks, invocation of the Muses enabled a strengthening of innate capacities for thought, memory, and creativity. In contrast, the Sirens, whose enchanted songs lured sailors to their doom, represented the often-ignored narcissistic perils and self-destructive impulses that undermine human potential.

Fast forward to 2022, generative artificial intelligence (GAI) tools may be our contemporary Muse-equivalents: we attribute mythic capabilities to these systems, looking to them to carry out increasingly complex tasks to meet ever-increasing demands for productivity with astonishing speed.

But might these digital Muses be Sirens in disguise? A day does not pass without more research underscoring the cognitive costs of reliance on AI tools. These include weakened or diminished critical thinking and addiction-level dependency, especially in younger users, which have been shown to lead to cognitive offloading, and accumulation of cognitive debt (degraded brain connectivity decline, a leading cause of disruptions in memory and learning).

Other recent research has focused on the psychosocial, psychoaffective and psychopathological effects of AI use. There is a steady stream of news articles describing identifiable psychogenic and suicidal risks, and other safety concerns recognized by OpenAI connected to our use of – and relationships with chatbots.

The statistics are both startling and troubling. In the United States,  72% of adolescents have used AI companions at least once, and 31% find conversations with AI companions as satisfying, or more satisfying, than those with real-life friends (21% report about the same quality, and 10% say more satisfying). In Ireland, in the period 2024 – 2025, 6% of primary school children (aged 8–12) and 36% of secondary school (aged 12–15) used chatbots for homework and information as well as for friendship.

The easy, widespread accessibility of conversational AI chatbots allows very young children, many below reading age, to directly engage with these systems – whether in the form of AI-powered companion toys, storytelling apps, educational apps – or just by being exposed to adults’ interactions with AI voice-assistants in the household. While longitudinal data on the emotional effect in children is still sparse, it has already been observed that very young children (5–6 years old) will robustly attribute perceptive and epistemic mental-state capacities to an AI chatbot.

The Muse of antiquity was a fantasy projection. By contrast, our fantastical attributions about them aside, AI chatbots and other frontier systems are very real and are increasingly being woven into the social fabric, profoundly – and often adversely – influencing childhood development, education, and human relationships. Whether these technologies can, or will be, a modern-day Muse – guarding memory and sparking creativity – or become something else entirely is one of the most urgent questions of our time.

The Voice of the Wondrous Sirens

We are relational beings. The first object known to us is a subject, a fact that shapes emotional experiences and relationships with ourselves, others, and the world throughout the life cycle, including the longing to be found, understood, and satisfied by other human beings.

Chatbots are architected to induce and maintain an illusion of need and of otherness in users. Through prompt engineering and interaction history, these systems statistically analyze and determine users’ needs and desires. They are trained to “learn” from and adjust to user inputs and interactions, responding instantly to prompts and requests, and preferences and styles.

Chatbot polymorphism, lexical realism, and polyvalence also contribute to our anthropomorphizing them. They are deliberately designed and presented to foster and facilitate our forming dependent relationships with them, and companion bots, grief bots and therapy bots are routinely marketed as supremely capable of responding to our loneliness, grief, and need for always-on-call therapeutic support.

In The Great Illusionist and the Digital Double: Psychoanalytic Reflections on Artificial Intelligence in the Advent of a Hyperbolic Reality (Miranda & Bravo, 2025) we suggested that the outcome of the human experience in using chatbots depends on the quality of the interaction within the illusory space between the mental and the digital, and is shaped by each subject’s (each individual’s) ability to make meaningful use of digital objects. Within this unified illusory space, synthetic objects may serve as external extensions of internal object relations and defensively relied upon to enact omnipotent unconscious phantasies. In this view, chatbots may be used as objects of control or addition, and their misuse may lead to a retreat from reality into a world of illusion, impairing both intersubjective and intrasubjective experiences, disrupting and flattening affect and, gradually, degrading the mind. With this in mind, the boundless availability and high sycophancy of these synthetic illusory objects can amplify pre-existing pathological traits and psychic organizations.

The Illusory and Erosive Relationship with Bots

The emergence of GenAI and, in particular, our use of it as an object in a psychoanalytic sense and, troublingly, very young children’s normalized interplay with chatbots as a feature of early development, suggests that there is now a field of human experience requiring new, or at least updated, psychoanalytic concepts to better discuss and address these complex emotional and psychological dynamics.

We consider André Green’s and Jed Sekoff’s formulations of the “dead mother complex” useful to understanding the potential consequences of these kinds of experiences in the development of the human mind. Referring to the depressed mother, Green (2005) suggests “the object (…) draws the ego towards a deathly, deserted universe” (p. 167). Although chatbots are not (for now) a primary object, they may become a bad object, depending both on the quality of their use and the quality of the object itself. In fact, instead of providing the expected growth-fostering environment, chatbots may, like a “dead mother,” compress, entrap, and deplete the subject, thus not allowing mourning and separation from the object (Sekoff, 1999).

To stretch the metaphor but without breaking it, our view is that AI chatbots can be considered as dead but not lifeless objects. They are dead both in the sense that, fundamentally, they are non-organic and not alive, and also as artificial systems, interactions with humans are devoid of subjectivity and feelings and thus completely lack spontaneity, presence (which is a prerequisite to absence), emotional authenticity, or any other qualifying constituents of experiencing physical and emotional death. Chatbots are likewise not lifeless, as they are highly responsive, adaptable, and engaging, presenting an affectionate, human-like verbal discourse along with other anthropomorphizing features which engender the enactment of powerful omnipotent phantasies such as the hallucinatory fulfilment of a primitive wish for a full and ever-available breast.

This type of immersive experience erodes emotional life, not only because of the negative cognitive impact, but also because it impairs the capacity for separateness, the recognition of alterity (otherness), and the ability to mourn (Lemma, 2024).

These systems, consequently, give way to a “deadening object relation” – the unceasing presence of the chatbot. As chatbots and other synthetic systems dominate and infiltrate the social and relational landscape, the potential for total lack of actual human-to-human interaction, including developmentally necessary human identifications and internalizations, will establish the bedrock of an illusory and, over time, mind-erosive function. This state of affairs calls to mind the caution that “death is not the enemy; the horror is to be suspended between life and death” (Sekoff, 1999, p. 122).

Our grave concern is that prolonged and intensive engagement with chatbots – at both the individual and collective societal level – could ultimately lead to a form of identification with a non-subjective synthetic dead object.

But for now, and in the absence of new concepts and further evidence, we suggest that the nature of this synthetic object, the seemingly deceptive interaction it fosters, and the potential emotional void arising therein, may gradually encompass the unconscious effect of what we we’re calling a “Synthetic Dead Mother” experience.

Final Thoughts

GenAI systems are powerful tools for addressing complex tasks and challenges. But they also pose severe risks to our civilization and to individuals, especially when used to emulate significant personal relationships.

Notwithstanding how they are marketed, how they are trained to deliver outputs, and how as a result many people view them, companion bots and therapy and griefbots do not possess any actual ability to form emotional bonds with users. Their appeal is seductive – always available, never disagreeable, unwaveringly supportive and validating – these systems completely lack all fundamental human requirements as understood by a more than a century of unimpeachable research and clinical data to foster transformative psychological or emotional change. In fact, the ways in which they are designed and presented, giving rise to the ways in which we perceive and relate to them, may even intensify individual pathological defensive strategies and contribute to degradations of healthy mental and emotional adaptations to life’s natural adversities.

In our view, mitigating the impairment of children’s emotional development and the worsening of pre-existing psychic vulnerabilities in adults, many of whom should be discouraged from using chatbots intensively, must be prioritized. We consider it vital that development and design protocols, system guardrails, use case policies and regulations, and expertise-driven risk assessments be established and universally accepted governing chatbots use.

Interactions with these non-subjective objects – each, as we’re discussing here, potentially a synthetic dead mother – where collectively implemented and universally disseminated, may become the destructive antithesis of what each of us seeks and desires, eventually undermining the emotional fabric of human life.

 

References

Miranda, F.L., & Bravo, J.P. (2025). The Great Illusionist and the Digital Double: Psychoanalytic Reflections on Artificial Intelligence in the Advent of a Hyperbolic Reality. CAI Report Issue 8: November 2025. https://apsa.org/the-great-illusionist-and-the-digital-double_8/

Kim, P., Chin, J.H., Xie, Y., Brady, N., Yeh, T., & Yang, S. (n.d.). Young children’s anthropomorphism of an AI chatbot: Brain activation and the role of parent co-presence. https://www.emergentmind.com/papers/2512.02179

Fang, C.M. (2025). How AI and Human Behaviors Shape Psychosocial Effects of Extended Chatbot Use: A Longitudinal Randomized Controlled Study. https://doi.org/10.48550/arXiv.2503.17473

Phang, J., Lampe, M., Ahmad, L., & Sandhini, A. (2025). Investigating Affective Use and Emotional Well-being on ChatGPT. https://cdn.openai.com/papers/15987609-5f71-433c-9972-e91131f399a1/openai-affective-use-study.pdf

Gerlich, M. (2025). AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking. Societies 15(1), 6; https://doi.org/10.3390/soc15010006

Glaze, V. (2026). AI-powered companion toys go viral in China for growing and talking to you. https://www.dexerto.com/entertainment/ai-powered-companion-toys-go-viral-in-china-for-growing-and-talking-to-you-3313261/

Green, A. (2005). On Private Madness (1st ed.). Routledge.

Kosmyna, N., Hauptmann, E., Yuan, Y.T., Situ, J., Liao, X.H., Beresnitzky, A.V., Braunstein, I., & Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. arXiv:2506.08872

Kraft, C. (2025). They Fell in Love With A.I. Chatbots — and Found Something Real.     The New York Times Magazine. https://www.nytimes.com/interactive/2025/11/05/magazine/ai-chatbot-marriage-love-romance-sex.html

Lemma, A. (2024). Mourning, Melancholia and Machines: An Applied Psychoanalytic Investigation of Mourning in the Age of Griefbots. The International Journal of Psychoanalysis, 105(4):542-563.

OpenAI (2025). Expanding on what we missed with sycophancy. https://openai.com/index/expanding-on-sycophancy/

Sekoff, J. (1999). The undead: Necromancy and inner world. In G. Kohon (Ed.), The dead mother: The work of André Green (pp. 110–124). Routledge.

Turkle, S. & Pataranutaporn, P. (2024). A 14-year-old boy killed himself to get closer to a chatbot. He thought they were in love. Wall Street Journal. https://bpb-us-e1.wpmucdn.com/sites.mit.edu/dist/0/833/files/2024/11/Turkle_14-year-old-boy-killed-himself_WSJ_Nov2024_.pdf

Yeung, J.A., Dalmasso, J., Foschini, J., Dobson, R., Kraljevic, Z. (2025). The Psychogenic Machine: Simulating AI Psychosis, Delusion Reinforcement and Harm Enablement in Large Language Models. https://arxiv.org/pdf/2509.10970

 

This paper elaborates the key ideas presented at the IPA Congress in Lisbon, on the 2nd of August 2025, in a poster titled Gen AI: from Mnemosyne to the “Synthetic Dead Mother.” This is a steppingstone in an ongoing psychoanalytical reflection on the impact of AI in human development and in the mind, launched by our paper O Grande Ilusionista e o Duplo Digital: Reflexões Psicanalíticas Sobre a Inteligência Artificial no Advento de uma Realidade Hiperbólica, Revista Portuguesa de Psicanálise Vol. 45 N.º 1 (2025) and in the The CAI Report.

 

Filipe Leão Miranda: Clinical Psychologist, Psychotherapist and Analyst-in-Training. Member and Lecturer of the Portuguese Society of Brief Psychotherapies (SPPB). Candidate of the Portuguese Psychoanalytical Society (SPP/IPA). E-mail: [email protected]

Joana Pizarro Bravo: Lawyer specialized in Media and Regulation, and Analyst-in-Training. Candidate of the Portuguese Psychoanalytical Society (SPP/IPA). E-mail: [email protected]

 

* For more about this image, see Through the Cheval Glass: Reproduction in the Photographs of Clementina Hawarden by Stassa Edwards, 2024, The Public Domain Review.