Technology Category
Machine Learning: Large Multimodal Models (LMMs)Natural Language Processing: (Large) Language ModelsCognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
A Disproof of Large Language Model Consciousness: The Necessity of Continual Learning for Consciousness
🤖 AI Summary
This paper addresses the fundamental question of whether contemporary large language models (LLMs) possess consciousness. Drawing on the falsifiability and nontriviality requirements of scientific theories of consciousness, we introduce the “Proximity Argument”: input–output equivalent systems—such as LLMs—fail to satisfy the formal constraints imposed by mainstream consciousness theories, and thus lack consciousness. Methodologically, we integrate formal logic, functional equivalence modeling, and an extended Unfolding/Substitution Argument. We further provide the first formal proof that continual learning is a necessary (though not sufficient) condition for human consciousness—a capacity inherently precluded in LLMs due to architectural limitations. Our analysis rigorously refutes consciousness in current LLMs, establishes continual learning as foundational to consciousness theory, and proposes the first theoretically rigorous and empirically actionable benchmark for AI consciousness assessment.
Technology Category
Application Category
📝 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 forms the basis of a disproof of contemporary LLM consciousness. I then show a positive result, which is that theories of consciousness based on (or requiring) continual learning do satisfy the stringent formal constraints for a theory of consciousness in humans. Intriguingly, this work supports a hypothesis: If continual learning is linked to consciousness in humans, the current limitations of LLMs (which do not continually learn) are intimately tied to their lack of consciousness.
Problem
Research questions and friction points this paper is trying to address.
Disproves consciousness in current Large Language Models
Shows continual learning is necessary for consciousness theories
Links LLMs' limitations to their lack of consciousness
Innovation
Methods, ideas, or system contributions that make the work stand out.
Using formal tools to analyze consciousness theories
Applying Proximity Argument to disprove LLM consciousness
Linking continual learning to consciousness in humans
🔎 Similar Papers
No similar papers found.
