Why “Is AI conscious?” resists a one-sentence answer
First distinguish intelligence, sentience, phenomenal consciousness, access consciousness, self-awareness, agency, and moral status.
Three evidence layers
1. Behavior
Self-reports, avoidance behavior, metacognition, and linguistic consistency.
2. Mechanism
Global sharing, recurrence, higher-order representations, self-models, and welfare-like representations.
3. Phenomenology
Qualia, sentience, and actual subjective experience: what it feels like.
Key point: Stronger evidence at layers 1 and 2 does not automatically establish layer 3.
The central divide: Substrate independence
Computational Functionalism
Prioritizes causal/information-processing roles over material. Artificial consciousness is possible in principle if the required structures can be implemented in silicon.
Biological Naturalism
Treats metabolism, homeostasis, and embodied self-maintenance as necessary life processes. Distinguishes functional simulation from experiential instantiation.
32 key terms
Search or filter by category. Keep function distinct from experience, and agents distinct from moral patients.
16 theories and current AI’s functional resemblance
Percentages below are qualitative functional-resemblance indicators for navigating this Atlas, not probabilities of consciousness.
How the question evolved
Five online ecosystems approach the same question differently
Separate current AI from future possibilities
Positions vary with time and context; read these as a map of debates.