The more precise version of "friendliness" is value alignment - Stuart Russell describes the problem well here[1], if you're interested. 1: http://edge.org/con

The more precise version of "friendliness" is value alignment - Stuart Russell describes the problem well here[1], if you're interested. 1: http://edge.org/conversation/the-myth-of-ai#26015 http://edge.org/conversation/the-myth-of-ai#26015