Nick Bostrom, IEET co-founder, talks about machine learning that can have the potential to self-improve. He talks about some catastrophic risks of such a complex system to humans. But what do we know? What can we predict? What should we assume when an artificial intelligence system improves upon itself and given freedom in the outside world? How does self-improving AI relate to the human mind? | Bostrom gave this talk at Winter Intelligence Oxford – AGI12 Conference.



Nick Bostrom, IEET co-founder, talks about machine learning that can have the potential to self-improve. He talks about some catastrophic risks of such a complex system to humans. But what do we know? What can we predict? What should we assume when an artificial intelligence system improves upon itself and given freedom in the outside world? How does self-improving AI relate to the human mind? | Bostrom gave this talk at Winter Intelligence Oxford – AGI12 Conference.



http://www.youtube.com/watch?v=uyxMzPWDxfI