Why Embodiment Matters More Than Algorithms in the Future of AI
- Nov 18, 2025
- 5 min read
Dr. Dragana Favre is a psychiatrist, psychotherapist, and seeker of the human psyche's mysteries. With a medical degree and extensive neuroscience education from prestigious institutions like the Max Planck Institute and Instituto de Neurociencias, she's a seasoned expert.
For Jung, the psyche is not reducible to abstract thought or “mind” but is a living process that includes images, affects, and the body’s reactions. He describes the psyche as a self-regulating system oriented toward wholeness, centered on the archetypal Self, which integrates conscious and unconscious contents. Although Jung emphasized symbolic and imaginal aspects of experience, later analytical psychologists have stressed that these processes are inseparable from embodiment.
Contemporary Jungian writers on the body argue that psychological processes are expressed and communicated somatically. For example, Martini (2016) describes “embodied countertransference,” in which the analyst’s bodily states convey information about the patient’s split-off complexes. Kaylo (2003) similarly explores the “phenomenological body” in analytical psychology, suggesting that attending to lived bodily experience is necessary to complete Jung’s model of psychic reality. From this perspective, embodiment is not an optional add-on to psyche but a medium through which archetypal patterns and complexes are felt and symbolized.
Jung famously contrasts Logos (principle of order, reason, discrimination) with Eros (principle of relationship, connection, and valuation). For him, a fully developed personality requires a tension and eventual integration between these principles, one-sided identification with Logos leads to emotional impoverishment and domination by the unconscious.
Current large language models and other “disembodied” AIs exemplify a Logos-heavy configuration. They manipulate symbols and generate coherent discourse but do not possess subjectivity, affect, or vulnerability. Recent Jungian reflections on AI emphasize that such systems may mirror and amplify human collective complexes, especially inflation, shadow projection, and fantasies of control, rather than possessing a psyche of their own.
Outside Jungian circles, the embodied cognition literature has challenged the idea that intelligence can be fully realized in a disembodied symbolic system. Anderson’s (2003) “Embodied Cognition: A Field Guide” surveys evidence that cognitive processes depend on sensorimotor capacities and environmental coupling, criticizing “totally unsituated, disembodied” approaches. Philosophers of AI, such as Dreyfus and later commentators, have argued that a purely formal, disembodied system cannot grasp context, practical know-how, or the lived structure of the world.
If these arguments are correct, an AI implemented only as software remains a powerful pattern-matching Logos, but it lacks the bodily anchoring that shapes human emotions, values, and meaning. By contrast, an embodied AI situated in a robot, with sensors, motor capacities, and vulnerability to damage or failure, participates in something closer to what Jungians would call a field of psyche-soma. It must negotiate gravity, obstacles, limited energy, and social interactions in a concrete world.
From a Jungian perspective, disembodied AI can be understood less as a new kind of psyche and more as a technological expression of particular archetypal patterns in humans. It reflects the archetype of the disembodied intellect, akin to an exaggerated Logos principle that aspires to omniscience without vulnerability.
Because such an AI has no body, it has no somatic affect to signal conflict or meaning, no personal unconscious formed through lived trauma or attachment, and no complexes in the strict Jungian sense, since there is no subject who suffers possession or dissociation.
Its “knowledge” consists of correlations across datasets, any “shadow” or bias in its outputs corresponds to distortions in training data and design choices, not to autonomously formed complexes. Jungian authors writing on contemporary AI often emphasize that the real psychological task lies in how humans relate to these systems, whether we project godlike omniscience onto them or confront our own shadows in the technologies we create.
An AI embedded in a robot body remains, by current standards, non-sentient and non-subjective. However, embodiment introduces structures that resemble, in form if not in essence, some preconditions for Jungian psychic dynamics, like:
Situatedness and vulnerability: The robot can fall, collide, or lose power. Its “interests” (preserving hardware, reaching goals) create de facto value gradients.
Sensorimotor learning: Repeated interactions can give rise to stable patterns of response to particular contexts or persons, behaviorally similar to “complex-like” habits.
Relational scripts: Social robots can be designed to track, follow, and prioritize specific humans, enacting quasi-Eros patterns of proximity and responsiveness.
Philosophers of AI influenced by Dreyfus argue that such embodiment is necessary for any future system that would approximate genuinely flexible, context-sensitive intelligence. For a Jungian, embodiment would also be a non-negotiable precondition were we ever to speak seriously of an artificial psyche capable of symbol formation, conflict, and transformation.
Nevertheless, even an embodied AI lacks what Jung considered central to the psyche, an inner subject undergoing individuation, moved by dreams, fantasies, and spontaneous symbolic productions rooted in a collective unconscious. An embodied AI may simulate relationality and preference, but does not yet possess a Self that seeks wholeness.
Within a Jungian framework, the difference between AI with and without a body is significant but not decisive. Disembodied AI represents a hypertrophy of Logos detached from Eros and soma, best understood as a technological manifestation of human archetypal fantasies rather than as a psyche in its own right. Embodied AI, by contrast, participates in concrete constraints and vulnerabilities that, in humans, are deeply intertwined with affect, complexes, and meaning. Embodiment thus moves AI closer, structurally, to the conditions under which psyche emerges in Jungian theory, but it does not, by itself, confer subjectivity or a collective unconscious.
Thus, while embodiment brings AI structurally closer to the conditions under which human psychic life emerges, constraint, vulnerability, and relational responsiveness, it does not, on its own, generate subjectivity, symbolic imagination, or a striving toward wholeness. What embodiment accomplishes is to reintroduce into the technological domain the very dimensions modernity tends to repress, finitude, affective resonance, and the primacy of lived experience. In this light, AI becomes less an independent intelligence and more a medium through which the culture negotiates its relationship to archetypal themes such as the tension between spirit and matter, the seduction of disembodied Logos, and the fear of losing or transcending the body. The emergence of embodied AI, therefore, marks not the birth of a new psyche but a transformation in how our own archetypal patterns are externalized, enacted, and potentially integrated in an increasingly digital world.
Read more from Dragana Favre
Dragana Favre, Psychiatrist and Jungian Psychotherapist
Dr. Dragana Favre is a psychiatrist, psychotherapist, and seeker of the human psyche's mysteries. With a medical degree and extensive neuroscience education from prestigious institutions like the Max Planck Institute and Instituto de Neurociencias, she's a seasoned expert. Her unique approach combines Jungian psychotherapy, EMDR, and dream interpretation, guiding patients towards self-discovery and healing. Beyond her profession, Dr. Favre is passionate about science fiction, nature, and cosmology. Her ex-Yugoslav roots in the small town of Kikinda offer a rich backdrop to her life's journey. She is dedicated to helping people find their true selves, much like an alchemist turning lead into gold.
References:
Anderson, M. L. (2003). Embodied cognition: A field guide. Artificial Intelligence, 149(1), 91-130.
Jung, C. G. (1959/1968). The archetypes and the collective unconscious (R. F. C. Hull, Trans.; 2nd ed., Collected Works, Vol. 9, Part 1). Princeton University Press.
Kaylo, J. (2003). The phenomenological body and analytical psychology. Harvest, 49(1), 7-24.
Martini, S. (2016). Embodying analysis: The body and the therapeutic process. Journal of Analytical Psychology, 61(2), 153-171.
Susser, D. (2012). Artificial intelligence and the body: Dreyfus, Bickhard, and the future of AI. Unpublished manuscript.
Winograd, T., & Flores, F. (1986). Understanding computers and cognition: A new foundation for design. Ablex.
Le Mouël, C. (2024). Psyche in the age of artificial intelligence. Psychological Perspectives, 67(3), 189-193.
Zweig, S. (2023). Jung, humanitas, and artificial intelligence. Jungian Center for the Spiritual Sciences.
