AI Consciousness Misattribution in High-Literacy Users: Observational Notes — LessWrong