I've wondered if the impression isn't so much caused by the answers themselves but a trust level has formed from its behavior. A core appeal of LLM bots is they

I've wondered if the impression isn't so much caused by the answers themselves but a trust level has formed from its behavior. A core appeal of LLM bots is they're instructed to be non-judgemental and ever-helpful, so in most scenarios one won't 'butt heads' with them. This obviously has some real positives for learning (if the responses could be accurate) since you're more likely to continue eagerly the less ashamed you are asking 'dumb' questions to more rapidly gain an understanding of something (anonymous accounts online similarly facilitate shame-free question asking). I wonder though if it might be having an effect where people begin to prefer AI over humans since with real humans, whether online or IRL, there's no consistency of response—people could be cranky, disinterested in responding or criticize someone for even asking. So the more one interacts with an ever-willing, ever-pleasant, non-judgemental, human-esque answer machine—even if it's hallucinating things—I could see how it could become more of a go-to and even a trust begin to grow from familiarity and its generally polite ability to listen.