Abstract

This article explores psychoanalytic perspectives, particularly Lacanian theory, on artificial intelligence, with a focus on chatbots, such as, ChatGPT. While ChatGPT is often viewed as mimicking human traits, such conceptions overlook the unique aspects of human subjectivity that AI lacks. Instead, psychoanalytic theory can reveal the social structures underlying AI and human interaction. We argue that ChatGPT should be seen as inherently relational, shaped by its developers and users, rather than as an independent, quasi-human agent. This relationality fosters a specific mode of sociality, analysed through Lacanian concepts, including the big Other, the five discourses, and the analyst-analysand relation. In doing so, we go beyond either rejecting the appearance of human qualities or embracing them as a new kind of AI subjectivity, be it technological or quasi-human.