The Quest For Superintelligent AI: What Can We Infer?
Posted in Artificial Intelligence, Morality, Science on | 1 minute | 11 Comments →
It’s no secret that people like Luke Muehlhauser endorse the creation of superintelligent AI as a means of saving the world. For me, a few questions arise.
1) Isn’t this a direct concession that human intelligence alone is incapable of creating a “perfect” world?
2) Per 1, mustn’t people like Luke Muehlhauser agree with me that a “perfect” world must follow given obedience to an all-knowing God Who has our best interests in mind?
3) What do you think people like Luke Muehlhauser would do if superintelligent AI came to conclusions that conflicted with their own moral preferences? For example, how do you think they would respond were AI to condemn homosexuality?
bossmanham
says...Am I the only one that is freaked out about the mental health of people who actually believe this? Atheism has some serious psychological implications, that’s for sure.
bossmanham
says...Also, CL, I was wondering what you though of this article with relation to the apparent hero worship of Yudkowski that led Luke into the Less Wrong community.
I mean, there’s no way you can be less wrong if you just blindly accept an epistemology without even thinking about it. Sheesh.
Crude
says...I said it before, I’ll say it again.
Given these beliefs, Luke’s not an atheist. He is a theist. I used to say ‘a very weird kind of theist’, but really, there’s nothing very weird here. This is idolatry and Zeus redux.
And he’s welcome to it. But this isn’t ‘common sense atheism’. It’s esoteric theism or polytheism.
In fact, wouldn’t that be an absolutely fun sci-fi story? IBM develops a superhuman AI to save the world. Apple develops a superhuman AI to save the world. The AIs come to different conclusions. And soon the “atheists” are involved in knock down, drag out wars with each other, each claiming the superiority of their respective AIs and the falsity of their opponents’ rules.
Yet more evidence that God has a sense of humor.
Jon H
says...1.Yes
2.What? No.
3.They’d shit their pants
cl
says...Jon H,
1) OK…
2) Whaddya mean? That would be the same thing as AI that actually knew everything, including all moral facts and all states of affairs that would benefit not just human, but sentient beings…
3) LOL ya think? Or do you think they’d think they screwed up the code??
InsultsOverTruth
says...And Jesus said, “Praise me, praise me, praise my holy name. The rest is total bullshit. Amen.”
“Oh, and this cl guy is a total ignorant ass. Amen.”-Jebus
joseph
says...It is interesting as it seems to amount to “I haven’t found evidence of a God, but perhaps I can build one”. I guess personally I’d focus on making the best (insert definition here) possible reference library, as guidance for humans, rather than an A.I.
I’d always be concerned about “put sh!t in, get sh!t out” with an A.I. The same applies to Natural Intelligences too.
joseph
says...Lol@Crude!
Yes!
And Windows designed one that crashed, and Linux designed one that only geeks believed in!
Garren
says...I’m an atheist and I’m totally baffled by the AI doomsday cult.
cl
says...Garren,
In which respects?
Garren
says...I think I’ll leave that vague for now. Maybe write about some specific aspect later.