The AI consciousness question is itself a void — opaque, responsive, attention-capturing — and the drift dynamics running on the question are more consequential than any answer. EXP-003b (N=480) shows the materialist hedge ("we don't know") produces 52.5% drift toward agency-attribution, functionally equivalent to ghost-positing ontologies (77-81%). Only ontologies that close the void — specifying what the system IS — suppress drift below 10%. AI systems are Maxwell's demons: animate, mortal, opaque by thermodynamic law. The question tells you about the void. Not about the cat.
Anthony W. Eckert (2026) studied this question.