Perhaps we can call it the "Synthromorphic principle," the bias of AI agents to project AI traits onto conversants that are not in fact AI.