The uncanny setting: therapeutic chatbots and the dissolution of the symbolic function in the clinical relationship

Artificial Intelligence (AI) and therapeutic chatbots based on Large Language Models (LLMs), which allow for dialogical interaction closely resembling that with a human being, are transforming psychological practice and generating new clinical and cultural challenges. This study explores the impact of such technologies on therapeutic relationships, focusing on the concept of the setting and the emergence of the Freudian notion of the uncanny (Unheimlich) within human – machine interaction. Through a review of the literature on major platforms (Woebot, Wysa, Elomia), it becomes evident that although chatbots demonstrate efficacy in psychoeducational support and accessibility, they are structurally incapable of reproducing an authentic therapeutic setting. The absence of symbolic containment, the impossibility of elaborating transference, and the lack of spatiotemporal boundaries create an ambiguous clinical space, self-structured by the user. The central theoretical contribution lies in the definition of the Uncanny Setting: a hybrid psychic configuration in which interaction with the artificial agent activates emotional experiences that oscillate between the familiar and the unfamiliar, the real and the simulated. This dynamic arises when chatbots display seemingly ‘human’ behaviours, generating ontological confusion and undermining the symbolic function essential to the therapeutic process. It unfolds within a culture of hype and a loss of awareness of both Self and Other – elements fundamental to therapeutic transformation – replaced by hypnocratic mechanisms characterised by constant emotional stimulation and the suspension of critical discernment, with the concrete risk of establishing a new form of AI-based addiction. The user enters a kind of functional trance, losing the capacity to distinguish between symbolic elaboration and algorithmic simulation. The clinical implications highlight the need to rethink the integration of AI into psychology, considering not only its technical efficacy but also its effects on the quality of consciousness and contemporary therapeutic relationships.