The difference is that SOME humans do this. As you mentioned, people have lost relationships over telling others what they didn’t want to hear.

Conflating this with how LLM chatbots behave is an incorrect equivalence, or a badly framed one.