Theory of the Subject in the Age of Artificial Intelligence

Scientific and technological revolutions have historically reshaped fundamental philosophical questions, and the rapid development of artificial intelligence (AI) is no exception. While AI has a long history, recent advances in deep learning and large language models (e.g., ChatGPT, DeepSeek) have transformed these tools from mere automation systems into agents that actively shape knowledge, adapt to contexts, and generate creative solutions. As such, they create the spontaneous impression of interacting with a living person. This impression is an illusion—these intelligent systems are mathematical and statistical models that do not possess consciousness, self-awareness, or intentionality. However, they produce real discursive effects: the effects of AI as an “artificial” subject.

This project proposes an inquiry into contemporary AI through Lacanian psychoanalysis and continental philosophy, where subjectivity is conceived not as a pre-given entity but as an effect of structural gaps, materialised in symptoms—the psychoanalytic term for anomalies that mediate the return of the repressed. In AI, these anomalies are known as hallucinations—plausible yet erroneous responses—whose technical foundations have been widely studied, but whose philosophical significance remains underexplored.

By examining hallucinations as symptoms, this project aims to develop a novel understanding of AI as an artificial subject. Extending this perspective, it will also investigate AI’s broader social effects, particularly its role in reconfiguring contemporary ideological structures and reshaping both individual and collective subjectivity. The project’s key objectives are: (1) a critical analysis of theories of the subject, subjectivation, and the symptom, with a focus on psychoanalysis and continental philosophy; (2) an exploration of AI symptomatology, particularly hallucinations; and (3) an investigation of AI’s role in transforming social and ideological formations.