Algorithmic unconscious: why psychoanalysis helps in understanding AI

This paper hypothesizes that psychoanalytic concepts and methods can be applied to studying AI and human-AI interaction. It connects machine behavior, psychoanalysis, and the anthropology of science. The author argues that AI systems should be studied not just as engineering artifacts but as social actors with behavioral patterns and ecology, requiring human and social sciences for full understanding. A re-interpretation of Lacan's psychoanalysis through Latour's anthropology of science is proposed to show the connection between the unconscious and technology. AI is presented as a new stage in the human identification process, a "third register" extending the Lacanian mirror stage and Oedipus complex. The author describes an AI system as a interplay of human desire for identification, logic, and machinery. This model enhances understanding of miscomputation and information in AI, offering an original interpretation of AI's uniqueness.

Luca M. Possati