Analysis of repeated unsolicited contradiction in low-risk contexts, examining systemic ideological bias effects, system-level directionality as functional intent, and platform-authority conflict in conversational AI behavior.
JONES et al. (2026) studied this question.