My assumption is OP doesn't know much about LLMs. Clearly speaking to chatGPT as a person (which I'm guilty of as well at times), but the personification goes further, as you're assigning it attributes that it can't have. Even in the comments people are drawing comparisons to a psychopath... It's just a bot. Also, their chatGPT's response style seems to me like it's been personalized. I also hate when chatGPT hallucinates and starts giving nonsense response, but I acknowledge its just "hallucinating"; what we don't want to start doing is attributing human emotions to a machine.