I'm dismissive of existential risks from AI because we 8 billion humans are destroying the planet by ourselves already. Rather than dealing with the existential risks we ourselves present to both other species and our own, we're pointing over there at something else that doesn't even exist yet [self-aware AI] and worrying about that
(
Read more... )
Comments 3
Reply
Yes, I agree, the danger of ever more powerful computers is what humans will either intentionally or unintentionally program them to do. There's zero chance of computers becoming self-aware.
The same human impulse to see gods and goddesses everywhere, in the sun, in trees, etc., is what compels people to see computers as self-aware. It is the mirror neurons in our brain that evolved as we became social animals. We can perceive each other as self-aware, so these mirror neurons overgeneralize and see all kinds of non-humans as self-aware also.
I don't know if people do this in every country, but in the US people will often re-hit a lit elevator button that they've already hit, thinking that hitting it again will make the elevator arrive faster. Elevator control systems are not programmed to arrive more quickly the more times you hit the button. But people instinctively presume that the elevator is self-aware and will respond to their expressed impatience. A computer chat program that mimics human expression easily tricks users ( ... )
Reply
But yeah, this is a phenomemon seen in many aspects. - That humans automatically think of it as self-evident that everything else is as conscious as them.
That's also how people interpret "signs" and patterns into things that tell them something about the status quo of life or about the future. Although there technically can't be something. Not only because no entity can know about that, but also because there's nothing out there that can read someone's mind. And that is omnipresent.
Reply
Leave a comment