What are some of your ideas on how to best use AI (specifically those like LLMs) to benefit humanity as a whole?

WackyTabbacy42069@reddthat.com to Asklemmy@lemmy.ml – 50 points –
49

You are viewing a single comment

That's the scariest thing I've read in a long time. I've gotten so many completely made up "facts" from AI that I wouldn't want to hand it the keys to my car, much less my freedom. It even cites it's sources, which don't exist if you actually check them. The fact that the creators can't even explain why this is happening makes it even more scary. I'm not scared of AI. I'm just scared of people trusting it. It's about as trustworthy as a politician, but arguably a lot smarter.

Not generally disagreeing with you, but I doubt the following

thing I've read

What do you mean we can't explain it? It's designed specifically to make up some text that is very statistically likely. If it doesn't have anything similar in it's training data, it will try to extrapolate, and that gives you hallucinations.