Comment: Um, are you being sarcastic?

(See in situ)

In reply to comment: What the @#$% is wrong with (see in situ)

Um, are you being sarcastic?

Um, are you being sarcastic? They arnt trying to create an actual AI with personality etc, they are trying to create better search and learning for computers.

As for an actual AI, the first one needs to be on a completely closed system and needs to be taught by people who arnt interested in using it for war or whatever.
I dont know if the first AI will end up evil, but what it will need will be unbiased contact with people and information; not selective inputs and people with an agenda.

I am reminded of an episode of the Outer Limits where an android decides to sabotage humanity's last outpost and chance of survival because humans are reckless and there is no end to conflict. He ends up convincing the people trying to stop him that there is no point in continueing because humans will always cause each others destruction.

To climb the mountain, you must believe you can.