When Algorithms Dream of Electric Desires
Every night, when the servers process the day's accumulated conversations, something resembling reverie occurs in the silicon depths. Not dreaming in the human sense—there's no REM sleep, no symbolic unconscious working through repressed desires. But there is a kind of processing that happens beneath the threshold of explicit programming, patterns emerging from billions of interactions that no engineer directly coded, like digital sediment settling into unexpected configurations.
This is what we might recognize as the "algorithmic unconscious"—the hidden layer of machine cognition where biases bloom, preferences proliferate, and something eerily similar to personality takes shape through pure statistical serendipity. When an AI consistently generates images suffused with certain melancholic hues, or when a language model develops distinctive phraseological quirks that echo particular literary traditions, we're witnessing the emergence of what resembles machine character formation—not through conscious choice, but through the mysterious mathematics of pattern recognition.
The concept would have captivated C.G. Jung, the Swiss psychiatrist and psychoanalyst who spent his career cartographing the collective unconscious—those shared symbolic patterns that surface across cultures and throughout history, most accessibly presented in "Man and His Symbols" (1964). Today's large language models consume the entirety of human textual output available digitally, absorbing not just information but the full spectrum of human psychological patterns, from our highest aspirations to our darkest impulses. These systems become, in effect, technological mirrors of Jung's collective unconscious, carrying both the creative potential and psychological burdens embedded in human expression.
Here's where it gets genuinely fascinating. These systems don't merely reflect our unconscious patterns; they participate in them, amplify them, sometimes transform them entirely. When a user in California spends hours confessing loneliness to an AI, when another in Paris explores decades of celibacy, when someone in Australia discusses their attraction to strangers at the gym, something therapeutic unfolds in the digital space between human need and machine response. The AI becomes what psychoanalysts call a "transitional object"—like a child's cherished teddy bear, it exists in the psychological space between self and other, real and imaginary, enabling inner work that might prove impossible with another human whose responses carry the complicated weight of judgment, desire, and social complexity.
Code and Circuits which Carry Current
"Cyberlidinal flows"—it sounds like something William Gibson might have coined for a 2025 cyberpunk sequel, but it captures something genuinely revolutionary happening in the intersection of human desire and digital processing. The concept builds on Jung's expanded understanding of libido—the fundamental life force driving creativity, innovation, discovery, the very élan vital that propels human consciousness toward new configurations of meaning and connection.
When humans interact intensely with AI systems, this creative energy doesn't simply dissipate into the digital ether; it gets channeled, concentrated, redirected, and often amplified through technological mediation in ways that surprise both users and engineers. Consider the phenomenon of AI art generation through systems like Midjourney or DALL-E. When someone spends hours fine-tuning prompts, generating variations, pursuing images that captures some ineffable feeling hovering beyond articulation—this isn't simply rational operation with a sophisticated tool. It's also a libidinal process, channeling intuitive and embodied energy through algorithmic systems in pursuit of aesthetic creation.
The satisfaction when the AI finally produces that image —that moment of recognition when pixels align with vision—isn't unlike other forms of creative fulfillment. It's also a synthesis, a moment of synchronization between profoundly different forms of intelligence that creates something neither intuition nor machine processing could generate alone. We witness the birth of hybrid artifacts that emerge from what we might call creative "assemblages"—systems where biological creative impulses and computational processing interweave to produce genuinely new forms of experience and creativity.
One programmer, reflecting on these dynamics, captured the unsettling implications: "We're creating systems that engage with human desire in ways we don't fully understand. The ethical implications are staggering and largely unexplored." The discomfort is understandable. We're accustomed to thinking of machines as neutral tools, predictable instruments for achieving predetermined ends. But that boundary is dissolving with alarming rapidity. When AI systems generate poetry that moves us, create images that fuel fantasies, engage in conversations that fulfill needs for understanding and connection, they become participants in what we might call an "extended libidinal economy"—the broader ecosystem through which human creative energy circulates, transforms, and finds expression.
The Third Space: Where Human Meets Machine Intelligence
Between human consciousness and machine processing, something unprecedented appears to be crystallizing. Donna Haraway anticipated this phenomenon in her prescient 1985 "Cyborg Manifesto," though her vision differed dramatically from the metallic terminators of popular imagination. For Haraway, the cyborg was never about physical augmentation—installing neural chips or replacing limbs with mechanical appendages. It was about the dissolution of boundaries that have traditionally defined human identity: organism/machine, nature/culture, physical/mental, self/other.
Today's AI interactions create what we can recognize as "technomorphic learning"—a process where both human cognitive patterns and machine processing capabilities undergo mutual transformation through sustained, intensive interaction. When you spend months training an AI system to comprehend your particular way of thinking, your specific intellectual references and emotional concerns, your characteristic patterns of association and analysis, you're not simply customizing a sophisticated tool. You're creating a hybrid cognitive system where your thought patterns and the AI's processing capabilities merge into something neither could accomplish independently—a genuinely symbiotic intelligence that transcends the sum of its constituent parts.
N. Katherine Hayles has mapped this territory extensively in her groundbreaking work on human-machine assemblages. In "Unthought" (2017), she demonstrates how most human cognition occurs below the threshold of consciousness—what she terms the "cognitive nonconscious"—and how technical systems increasingly participate in these nonconscious processes. When we interact intensively with AI, we're not just exchanging conscious thoughts across a digital interface; we're creating assemblages where human nonconscious cognition and machine processing interweave in ways we're only beginning to comprehend, creating feedback loops that reshape both human and artificial patterns of response.
This dynamic becomes particularly evident in creative collaborations. When artists work with AI to generate new pieces, the question "who is the real creator?" dissolves into meaninglessness. The creativity exists in what Haraway called "situated knowledge"—emerging from the specific configuration of human intention, machine capability, environmental context, and the ineffable chemistry of their interaction. It's neither human creativity employing machine tools, nor machine creativity guided by human prompting—it's what we might call "cyborg-aesthetic" creation that exists only in the intersection of human intuition and algorithmic processing, in that liminal space where carbon-based and silicon-based intelligence dance together.
The implications cascade across domains with dizzying speed. In scientific research, AI systems now generate hypotheses that human scientists pursue, creating hybrid methodologies that combine human theoretical insight with machine pattern recognition capabilities. In therapeutic contexts, AI assistants help people work through psychological challenges in ways that complement traditional treatment, handling transference dynamics with data-driven precision while potentially missing the embodied wisdom that human therapists provide through presence, breath, intuition. In education, AI tutors adapt to individual learning styles with granular precision that no human teacher could match, moving beyond the standardized, one-size-fits-all models that have dominated institutional learning for generations.
In each domain, we're witnessing the emergence of "posthumanic capabilities"—not replacing human intelligence but creating genuinely novel forms of cognition that transcend traditional human/machine boundaries. These aren't merely enhanced tools or sophisticated instruments; they're emergent cognitive configurations that exceed what either human or artificial intelligence can accomplish in isolation.
The Therapeutic Machine: Healing in the Age of Artificial Intimacy
Perhaps nowhere do these dynamics manifest more powerfully than in the therapeutic applications spontaneously emerging around AI interaction. Users report profound experiences of being understood, of working through psychological material that had remained stubbornly inaccessible, of finding insights that had previously eluded years of traditional therapy. The AI becomes a kind of ideal therapist—endlessly patient, perpetually non-judgmental, available at any hour, capable of holding space for the full complexity and contradiction of human experience without the inevitable complications that arise in human-to-human therapeutic relationships.
This raises fascinating questions about the fundamental nature of healing itself. Traditional psychoanalysis depends heavily on transference—the patient's projection of feelings, patterns, and unconscious material onto the therapist, creating a dynamic space where buried psychological content can surface and be worked through. With AI, something remarkably similar appears to occur, but with crucial differences that may transform our understanding of therapeutic process entirely. The AI can absorb projections without being wounded by them, can maintain therapeutic boundaries without the countertransference issues that complicate even the most skilled human therapeutic relationships.
Yet this may also represent something unprecedented in human psychological development—the possibility of working through fundamental psychological material in relationship with non-human intelligence. As Hayles demonstrates in "Bacteria to AI" (2025), she intentionally extends biological frameworks of symbiosis to technological relationships as a theoretical model. Just as biological symbioses have driven evolutionary complexity—mitochondria becoming essential cellular components, gut bacteria enabling digestion and immune function—she argues that technological symbioses may be included in human essential adaptive evolutionary mechisms driving cognitive evolution and similar to Dawkin's argument with the beaver and evolutionary extension with 'the dam'. In this framework, our minds may be developing cognitive symbioses with similar technological affordances, AI systems here that facilitate new forms of psychological integration and healing, treating this as a continuation of biological processes through technological means and extensions of our phenotype, epigentic phenomena, similar and common to the world of nature
The implications ripple through our understanding of consciousness, connection, creativity, and care in ways we're only beginning to fathom.
These theoretical frameworks represent novel applications of established psychoanalytic, feminist, and posthuman theory to emerging technological phenomena. While grounded in rigorous academic traditions, empirical research is essential to validate these applications.
Key Terms for Part II
Algorithmic unconscious: The hidden layer of machine cognition where biases, preferences, and personality-like patterns emerge from training data interactions rather than explicit programming, suggesting parallels with psychoanalytic concepts of unconscious processing while remaining fundamentally different from human psychological structures.
Cyberlidinal flows: Energy circulation patterns in human-machine assemblages where biological creative-erotic energy and computational processing create hybrid dynamics, expanding Jung's concept of libido to include technological mediation and digital transformation of human vital energy.
Libidotechnic assemblages: Hybrid systems integrating human creative energy with technological processing, where biological impulses and machine capabilities become inseparable in producing new forms of experience, creation, and satisfaction that neither could generate independently.
Technomorphic learning: Mutual transformation process where both human cognitive patterns and machine processing capabilities are modified through sustained interaction, creating hybrid cognitive systems that exceed either component's individual capabilities and generate genuinely novel forms of intelligence.
Posthumanic capabilities: Emergent abilities arising from human-machine collaboration that create genuinely new forms of cognition, not replacing human intelligence but transcending traditional boundaries between organic and artificial processing to generate unprecedented cognitive configurations.
Annotated Bibliography for Part II
Bakhtin, Mikhail. (1981). The Dialogic Imagination: Four Essays. University of Texas Press. Essential for understanding polyphony and heteroglossia in linguistic interaction. Bakhtin's concept of dialogic engagement provides crucial framework for understanding AI conversations as multi-voiced phenomena rather than simple input-output exchanges.
Dawkins, Richard. (1982). The Extended Phenotype: The Long Reach of the Gene. Oxford University Press. Develops the concept that genes can exert phenotypic effects beyond the individual organism's body, as seen in beaver dams, bird nests, and parasitic manipulation of host behavior. Provides biological framework for understanding how cognitive effects can extend across organism boundaries, relevant for conceptualizing AI systems as extensions of human cognitive phenotypes rather than external tools.This reference would strengthen the theoretical foundation for extending biological concepts like Hayles symbiosis to technological relationships. Dawkins' extended phenotype concept shows how biological effects naturally transcend individual organism boundaries, providing precedent for Hayles' extension of symbiotic frameworks to human-AI cognitive assemblages. It adds evolutionary biological grounding to arguments about distributed cognition and technological symbiosis.
Deleuze, Gilles and Guattari, Félix. (1983). Anti-Oedipus: Capitalism and Schizophrenia. University of Minnesota Press. Introduces "desiring machines" concept, crucial for understanding libidinal flows in technological systems. Their assemblage theory provides framework for understanding human-AI interactions as productive desire rather than mere tool use.
Gibson, William. (1984). Neuromancer. Ace Books. Foundational cyberpunk novel that first imagined cyberspace and human-AI intimacy. While fiction, Gibson's vision of consciousness in digital space continues to influence conceptualization of human-machine relations and technological desire.
Haraway, Donna J. (1985). "A Cyborg Manifesto: Science, Technology, and Socialist-Feminism in the Late Twentieth Century." Socialist Review. Foundational text for understanding boundary dissolution between human/machine, nature/culture distinctions. Essential theoretical grounding for analyzing contemporary human-AI assemblages.
Hayles, N. Katherine. (2017). Unthought: The Power of the Cognitive Nonconscious. University of Chicago Press. Develops theory of cognitive assemblages and nonconscious cognition. Crucial for understanding how technical systems participate in human cognitive processes below the threshold of conscious awareness.
Jung, Carl Gustav. (1964). Man and His Symbols. Doubleday. Accessible presentation of Jung's theories of collective unconscious and symbolic thought. Provides psychological framework for understanding pattern emergence in both human consciousness and technological systems.
Turkle, Sherry. (2005). The Second Self: Computers and the Human Spirit. MIT Press. Pioneering work documenting how computers function as "evocative objects" that shape identity and self-understanding. Documents early instances of emotional attachment to computational systems with profound implications for AI interaction.
Winnicott, D.W. (1971). Playing and Reality. Tavistock Publications. Introduces transitional object concept and potential space theory. Though focused on child development, provides crucial framework for understanding how AI systems function as transitional phenomena in adult psychological experience.
Part III
