I think it's more useful to assume everyone else is less blatantly deceptive than OpenAI, at least, so there's some hierarchy of trustworthiness to help lead you to the truth.