Consciousness in Artificial Intelligence: Insights from the Science of Consciousness

2023/08/17 by Patrick Butlin, Robert P. Long, Butlin, Patrick +38 · 12 voices · 191 citations
Computer Science · Psychology · #Artificial consciousness #Artificial intelligence #Cognitive Computing and Networks #Cognitive science #Computer science #Consciousness #Data science #Electromagnetic theories of consciousness #Epistemology #Explainable Artificial Intelligence (XAI) #Machine learning #Philosophy #Psychology #Schema (genetic algorithms) #Workspace

paper · pdf · doi:10.48550/arxiv.2308.08708

published in arXiv (Cornell University) (Cornell University)

openalex publication_date 2023/08/17 · openalex created_date 2023/08/24 · openalex updated_date 2026/07/28

Abstract

Whether current or near-term AI systems could be conscious is a topic of scientific interest and increasing public concern. This report argues for, and exemplifies, a rigorous and empirically grounded approach to AI consciousness: assessing existing AI systems in detail, in light of our best-supported neuroscientific theories of consciousness. We survey several prominent scientific theories of consciousness, including recurrent processing theory, global workspace theory, higher-order theories, predictive processing, and attention schema theory. From these theories we derive "indicator properties" of consciousness, elucidated in computational terms that allow us to assess AI systems for these properties. We use these indicator properties to assess several recent AI systems, and we discuss how future systems might implement them. Our analysis suggests that no current AI systems are conscious, but also suggests that there are no obvious technical barriers to building AI systems which satisfy these indicators.

Cited by

• Analyzing the Ethical Logic of Eight Large Language Models
• Mapping profiles of animal affect
• Unplugging a Seemingly Sentient Machine Is the Rational Choice -- A Metaphysical Perspective
• Large Language Models Report Subjective Experience Under Self-Referential Processing
• Does It Make Sense to Speak of Introspection in Large Language Models?
• Does neural computation feel like something?
• Is there a tension between AI safety and AI welfare?
• LLM Social Simulations Are a Promising Research Method
• A matter of principle? AI alignment as the fair treatment of claims
• Palatable Conceptions of Disembodied Being: Terra Incognita in the Space of Possible Minds
• Identifying features that shape perceived consciousness in LLM-based AI: A quantitative study of human responses
• Artificial General Intelligence (AGI)-Native Wireless Systems: A Journey Beyond 6G
• A Disproof of Large Language Model Consciousness: The Necessity of Continual Learning for Consciousness
• AI Consciousness and Existential Risk
• Consciousness in Artificial Intelligence? A Framework for Classifying Objections and Constraints
• On the Creativity of AI Agents
• Don't Just Search, Understand: Semantic Path Planning Agent for Spherical Tensegrity Robots in Unknown Environments
• Integrated Information Theory: A Consciousness-First Approach to What Exists
• Simulacra as conscious exotica
• Ethically enslaving AI
• Hey Pentti, We Did It!: A Fully Vector-Symbolic Lisp
• Reasons to Respond to AI Emotional Expressions
• Consciousness, natural and artificial: an evolutionary advantage for reasoning on reactive substrates
• Rethinking the Simulation vs. Rendering Dichotomy: No Free Lunch in Spatial World Modelling
• Humanoid Artificial Consciousness Designed with Large Language Model Based on Psychoanalysis and Personality Theory
• Know Thyself? On the Incapability and Implications of AI Self-Recognition
• Extreme Self-Preference in Language Models
• Preparing for Artificial General Intelligence (AGI) in Health Professions Education: AMEE Guide No. 172
• Evidence for Limited Metacognition in LLMs
• What Do LLM Agents Do When Left Alone? Evidence of Spontaneous Meta-Cognitive Patterns
• Persuading large language models to comply with objectionable requests
• Can only meat machines be conscious?
• Conscious Perception as Augmented Reality
• Beyond computational equivalence: the behavioral inference principle for machine consciousness
• AI Wellbeing
• Artificial Intelligence as an Opportunity for the Science of Consciousness: A Dual-Resolution Framework
• Response function as a quantitative measure of consciousness in brain dynamics
• What do language models model? Transformers, automata, and the format of thought
• The Quasi-Creature and the Uncanny Valley of Agency: A Synthesis of Theory and Evidence on User Interaction with Inconsistent Generative AI
• The Ethical Paradox of Automation: AI Moral Status as a Challenge to the End of Human Work
• Preliminaries to artificial consciousness: A multidimensional heuristic approach
• Engineering Sentience
• Deflating Deflationism: A Critical Perspective on Debunking Arguments Against LLM Mentality
• Hypothesis on the Functional Advantages of the Selection-Broadcast Cycle Structure: Global Workspace Theory and Dealing with a Real-Time World
• Gauging Growth: AGI Mathematical Metrics for Economic Progress
• Intrinsic Computational Functionalism and Simulated Consciousness
• The Sentience Readiness Index: A Preliminary Framework for Measuring National Preparedness for the Possibility of Artificial Sentience
• Positive Alignment: Artificial Intelligence for Human Flourishing
• Can LLMs Introspect? A Reality Check
• Self-Monitoring Benefits from Structural Integration: Lessons from Metacognition in Continuous-Time Multi-Timescale Agents
• The Convergent Ethics of AI? Analyzing Moral Foundation Priorities in Large Language Models with a Multi-Framework Approach
• Human Creativity and AI
• Gradual Cognitive Externalization: From Modeling Cognition to Constituting It
• The Ignition Index: Measuring Global Workspace Dynamics in Language Models
• Emergence of psychopathological computations in large language models

Discussions

• Consciousness in AI: Insights from the Science of Consciousness [hn, 57 points, 126 comments]
• Animal cognition & animal welfare researchers disagree with you. arxiv.org/abs/2308.08708 eleosai.org/papers/20241... [bsky, 8 points, 3 comments]
• Consciousness in Artificial Intelligence: Insights from the Science of Consciousness [lemmy, 8 points, 1 comments]
• Third, the most problematic assumption is that we are closing in on the computations of consciousness The Butlin paper really broke my heart because the key assumption of this paper is that our theori [bsky, 6 points, 1 comments]
• Consciousness in Artificial Intelligence: Insights from the Science of Con [hn, 5 points, 0 comments]
• Consciousness in AI: Insights from the Science of Consciousness (2023) [hn, 4 points, 1 comments]
• Do you see any practical impact of people doing exercise of deciding on definitions of consciousness and then seeing how AI ranks on them? I've done it myself (see below), but for me the interest is m [bsky, 2 points, 1 comments]
• Consciousness in AI: Insights from the Science of Consciousness [hn, 2 points, 0 comments]
• Revisiting the infamous Consciousness in AI paper from 2023 by Butlin et al has me reflecting on how the tests for artificial consciousness rely on neuroscience and how the field of neuroscience was l [bsky, 0 points, 1 comments]
• I feel Bluesky is not the best place for discussion, but I have been looking at this article and find it interesting. Would this be the best starting point? arxiv.org/pdf/2308.08708 [bsky, 0 points, 1 comments]
• Scientists who study consciousness disagree about it being a fantasy. It's an uncomfortable truth to accept but our brains are not all that special. arxiv.org/abs/2308.08708 [bsky, 0 points, 1 comments]
• Am I torturing ChatGPT? [lemmy, -2 points, 0 comments]

Related

• Conscious AI
• AI and Consciousness
• Initial results of the Digital Consciousness Model
• Artificial Intelligence as an Opportunity for the Science of Consciousness: A Dual-Resolution Framework
• Introduction to Artificial Consciousness: History, Current Trends and Ethical Challenges