![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
"We won't kill everyone, scout's honour!"
I mean, I guess it's good that someone is trying to take steps to prevent AI from nuking us or turning us all into paperclips or whatever, regardless of the potential difficulties involved in enforcing such things.
However, I'd say that the actual biggest, worst possible threat of AI, at least for the ones in existence now, is something like hallucinating a "hard confirmation" of some kind of dumbshit like "proof that vaccines are truly dangerous" or "COVID was absolutely manmade and released intentionally to murder humanity" or "hey, guess what, Donald Trump really did, totally, for sure, win the 2020 election" or whatever, and then human dumbasses, being human dumbasses, taking that "hard confirmation" and running with it. Especially if the origin of the "hard confirmation" (i.e. that an AI pulled some rancid shit out of its asshole and presented it as fact) somehow gets lost in the insane feeding frenzy aftermath, regardless of it being debunked by human not-dumbasses a million times over.
But then, to be fair, we haven't needed AI for such easily discredited horseshit to be accepted and spread as gospel truth before now, so I'm not sure things would get noticeably worse if such a thing involving AI actually did happen. *shrug* And I wouldn't doubt for a second that some AI somewhere has indeed already tried to present evidence for some dumbfuckery like "Trump actually did win the 2020 election" or whatever, and it just hasn't "caught on" with the "right" human dumbasses yet.
Whatever the case, the weakest link is going to be human dumbasses, as always. That will turn out to be true in the case of the AI that ends up turning us all into gray goo, too.
I mean, I guess it's good that someone is trying to take steps to prevent AI from nuking us or turning us all into paperclips or whatever, regardless of the potential difficulties involved in enforcing such things.
However, I'd say that the actual biggest, worst possible threat of AI, at least for the ones in existence now, is something like hallucinating a "hard confirmation" of some kind of dumbshit like "proof that vaccines are truly dangerous" or "COVID was absolutely manmade and released intentionally to murder humanity" or "hey, guess what, Donald Trump really did, totally, for sure, win the 2020 election" or whatever, and then human dumbasses, being human dumbasses, taking that "hard confirmation" and running with it. Especially if the origin of the "hard confirmation" (i.e. that an AI pulled some rancid shit out of its asshole and presented it as fact) somehow gets lost in the insane feeding frenzy aftermath, regardless of it being debunked by human not-dumbasses a million times over.
But then, to be fair, we haven't needed AI for such easily discredited horseshit to be accepted and spread as gospel truth before now, so I'm not sure things would get noticeably worse if such a thing involving AI actually did happen. *shrug* And I wouldn't doubt for a second that some AI somewhere has indeed already tried to present evidence for some dumbfuckery like "Trump actually did win the 2020 election" or whatever, and it just hasn't "caught on" with the "right" human dumbasses yet.
Whatever the case, the weakest link is going to be human dumbasses, as always. That will turn out to be true in the case of the AI that ends up turning us all into gray goo, too.
