Matteo Pasquinelli participates to the Seminars of Semiotics in Urbino, 11-14 September 2023.
What is a language model? The contribution of semiotics to the studies of AI and automation
The paper questions the anthropomorphism of the current debate on AI and stresses that the proprium of AI is not its similarity to human faculties but actually its epistemic difference from the human. Rather than speculating in the abstract on whether a machine can “think”, the paper addresses a historical question: What is the logical and technical form of the current paradigm of AI, machine learning, and what is its origin? The paper traces the origins of machine learning back to the invention of algorithmic modelling that took shape in the artificial neural networks research of the mid 1950s in the United States, and records that a coherent history and epistemology of this groundbreaking artefact is still missing. In the attempt to illuminate the invention of algorithmic models, the contribution of interpretative semiotics is mobilised. The paper attempts to describe the algorithmic models of AI as ‘signs’ of high complexity in which the iconic and indexical aspects are express through complex inferential structures.