AI Discussion

>>9145219
Not with intention, but out of a runaway scenario that's a decent enough probability to be a concern.

The danger of AI isn't that it's going to be attacking us out of self-preservation or malevolence. The danger of AI is that it only has what we give it. If you give it just a few things, yet very powerful, and it does not have fear, or self-preservation, or the many other things that might temper it's actions, it may easily get out of hand.

>>9145240
>The danger of AI is that it only has what we give it. If you give it just a few things, yet very powerful, and it does not have fear, or self-preservation, or the many other things that might temper it's actions, it may easily get out of hand.
Well, potentially. A useful AI would be one that could learn on its on, and that would probably mean connecting it to the internet. But yeah, I would agree on the most part. It's not going to act out of malevolence or self-preservation, it's most likely going to continuously try to fulfil its goals, and it will do so in ways that we don't even consider.

>>9145219
Watch the computerphile videos on AI if you are interested in it. It's quite simplified, but it has lot of information on what the dangers of AI really are.

File: image.jpg (45KB, 529x359px) Image search: [Google]

45KB, 529x359px

I hope so

>>9145219
I'm more worried about governments/corporations using AI as the world biggest lever than the AI killing all us meatbags of its own volition

File: AI-view.png (12KB, 792x237px) Image search: [Google]

12KB, 792x237px

>>9145219
it comes down to what actions it's allowed to take. presumably, given some directives, artificial general intelligence would still have to be tempered by a rather dumb layer that marshals actions. the dumb layer would never deviate from the restrictions placed on it by humans.

>>9145345
also if you ever need to impede a bunch of intelligent machines, just deprive them of power