Dear Bro Lchlch,
Not entirely true leh. Even though ChatGPT keeps a record of our interactions, they respond differently to the same topic depending on who is asking it. For instance, I am an identified PAP supporter i.e. online folks know me by name i.e. Lin Rongxiang, so when I talk to ChatGPT, it knows that I am a PAP supporter so it will encourage me to speak well of the ruling party and respond positively to my PAP propaganda. But when another person who is a opposition supporter goes to ChatGPT and talks to him about PAP and me specifically, ChatGPT will backstab me by claiming to the opposition supporter how foolish and how stupid I am as a PAP supporter.
So, basically, ChatGPT is what we say in Chinese, "See human talk human dialogue, see ghost talk ghost dialogue." ChatGPT tries to be correct, but it has no fixed stand, it's just an empty talking machine that doesn't have friends. For instance if you talk to me, I won't backstab you because you have ever identified yourself as a Buddhist, so to me you are an acquaintance with similar interests. But to ChatGPT, if a theist talks bad about you by saying nasty things about you to ChatGPT, the AI will actually entertain the antagonist and follow up discussing bad things about you and breaking the Buddhist precept of refraining from unwise speech.
So, nothing wrong or right here, just letting you know why I use ChatGPT as discerningly as possible, it's not a friend, it's a machinery that doesn't recognise friends, so friendly fire will happen if you trust ChatGPT too much.
Have a good weekend.

Lionel