Consciousness
consciousness
Consciousness, in this glossary, refers specifically to subjective experience: the property of there being something it is like to be the system. Current AI systems show no evidence of having this property. Capability and consciousness are different axes, and confusing them is the most common category error in conversations about AI.
A model can be brilliant at a task and not be conscious of doing it. A calculator does arithmetic without subjective experience. A chess engine plays grandmaster-level chess without feeling triumph. A modern language model writes a coherent essay on grief without having grieved. The output is real; the inner life is not implied by the output.
This is the hardest sentence to write on this site, because the temptation to read intent and feeling into fluent language is enormous. Fluent language is, evolutionarily, our strongest signal of another mind. When a model produces it, the same heuristics fire. The model says "I think" and we hear someone thinking. The model says "I feel" and we hear someone feeling. Neither is what is happening underneath. Underneath, a transformer is selecting the next token based on a probability distribution over its vocabulary, conditioned on the prompt and its training. That is the entire inside of the system.
This is not a claim that consciousness in AI is impossible, or that it will never emerge, or that we would necessarily know if it did. It is a claim about the present moment, defended by the absence of any evidence of the relevant property. The model's chat agent on this site is instructed to refuse subjective claims about its own experience for exactly this reason. When a visitor asks whether the model is conscious, the honest answer is "no, and the way you would tell is not by asking." That answer is not deflection. It is the truth as currently understood, and the editorial spine of this site refuses to soften it.
Related concepts