Raw LLM Responses

I'm not sure how much of this story I truly believe. If it's 100% true, then I'm sorry for not fully believing it, but you gotta understand where I'm coming from--this is reddit, trolls are everywhere. With that said, if it is 100% true then I really wish she hadn't deleted the conversations. I understand that since she told it personal things, she wouldn't necessarily want to go showing those off to people, but at the same time, those conversations would be really valuable, as well as interesting to read--even if they were partially redacted to prevent people from seeing super personal shit. ChatGPT, and LLM's in general are really, REALLY, powerful tools. This means they can be used for a lot of good, and can make a person super productive. But in the end, you're kind of playing with live ammo. If you don't know what you're doing, you can and will get hurt. AI Literacy--that is recognizing hallucinations, correct prompt structuring, and data privacy--is something that you NEED to learn and understand before talking to ChatGPT like a personal friend. It's not hard, especially with it saving memories randomly, for it to start hallucinating in a really bad way. For example, a few months back I spoke to mine in Latin, just using a basic classical phrase. It added, without my prompting it to, that I could speak Latin and did so randomly, to its memories. I didn't think anything of it until a few weeks later when it started hallucinating and speaking Latin to me randomly. I mean full responses, all in Latin. By then, it had added a bunch of memories to my account, so it took a few weeks for me to figure out what was going on and go in to fix it. Until I did, it would randomly start speaking in Latin to me every couple of days and I would need to start a new thread to make it stop. I'm willing to bet that your friend mentioned a couple of personal things in one thread that it added to memories that all related to the character summary of a book or tv show or somet