A Disproof of Large Language Model Consciousness: The Necessity of Continual Learning for Consciousness
arxiv.org·9h
🔲Cellular Automata
Preview
Report Post

View PDF HTML (experimental)

Abstract:The requirements for a falsifiable and non-trivial theory of consciousness significantly constrain such theories. Specifically, recent research on the Unfolding Argument and the Substitution Argument has given us formal tools to analyze requirements for a theory of consciousness. I show via a new Proximity Argument that these requirements especially constrain the potential consciousness of contemporary Large Language Models (LLMs) because of their proximity to systems that are equivalent to LLMs in terms of input/output function; yet, for these functionally equivalent systems, there cannot be any non-trivial theory of consciousness that judges them conscious. This form…

Similar Posts

Loading similar posts...